[Server-devel] How can I delete 10 years of git history in the XSCE repo?

[George Hunt]

The XSCE repo, inherited from laptop.org has grown to 80MB+.
Recently, then the raspberry pi3 came out, as an experiment, I started
a branch off or XSCE, which was initially aimed at a quick and dirty
demo that our code could run on top of raspbian.

Later, when that demonstration, seemed successful, I circled back and
modified the new branch so that it would also work on top of CentOS,
and Debian.

Now, I'd like to chop off the history at the point where of our last
major release, and create a new truncated repo.

I found a recipe in the middle of the page at:
http://stackoverflow.com/questions/4515580/how-do-i-remove-the-old-history-from-a-git-repository

#!/bin/bash
git checkout --orphan temp 
git commit -m "Truncated history"
git rebase --onto temp $1 
git branch -D temp

What I encountered was a huge number of conflicts, during the rebase step
which I did not understand, and seemed too error prone to continue

What's wanted is a repo that includes all of the commits starting from
https:/github.com/xsce/xsce --branch release-6.0 and including all of
the commits on https://github.com/georgejhunt/xsce --branch base-6.2.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] CentOS 7 quirk: VGA cable not re-attachable

[George Hunt]

I do most of my testing of real hardware, and came across the VGA problem
also,  I've migrated to using the hdmi input to my monitor, and dongles
that translate to hdmi.  And I'm able to swap the monitor cable
indiscriminately.

On Mon, Aug 29, 2016 at 3:59 PM, Adam Holt  wrote:

> On Mon, Aug 29, 2016 at 5:50 PM, James Cameron  wrote:
>
>> Only tested one monitor?  Try another; the EDID communication path
>> over VGA could have problems, or the data in monitor Flash unusual.
>>
>
> I tried a different brand of LCD monitor, and the problem remains:
>
> Fedora 22 NUC allows re-attaching of the VGA cable every time, whereas
> both CentOS NUC's are highly intermittent (the VGA cable fails to show any
> video upon re-attachment, the vast majority of the time).
>
> In hindsight, the other LCD monitor (tested across all 3 NUC's) behaves
> very much the same way.  Perhaps 10% of the time (or less) video
> re-appears, sometimes triggered by starting/leaving X Windows using
> Ctrl-Alt-F1 or Ctrl-Alt-F2 (or running "reboot" from ssh, on any machine
> that was running X Windows at the time, causing video to suddenly re-appear
> during the shutdown process) but regrettably on both CentOS 7 NUC's
> (NUC6i3SYH & D34010WYKH) I haven't found any repeatable pattern that
> permits video to be usefully restored (without a full reboot!)
>
> PS Problem occurs independent of CentOS 7 NUC's that are running X Windows
> and those that are not.  I mention X Windows only because leaving X /
> re-entering X seems to (sometimes) facilitate recovery of the video signal
> in certain not-quite-repeatable situations.
>
> Also check age of monitor.  Mix not old and new wineskins.
>>
>
> The earlier test monitor's from the prior decade it's true, but this newer
> test from a much more recent vintage ;)
>
> --
>> James Cameron
>> http://quozl.netrek.org/
>> ___
>> Server-devel mailing list
>> Server-devel@lists.laptop.org
>> http://lists.laptop.org/listinfo/server-devel
>>
>> --
>> 
>> 
>> Unsung Heroes of OLPC, interviewed live @
>> http://unleashkids.org !
>>
>
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Fedora 24 released / Fedora 22 "end-of-life" July 19 2016

[George Hunt]

Peter,

What can you tell me about rpi3 bluetooth support in the F-24 images that
will be coming out soon?

By necessity, I've been playing with raspian, because there's been a lot of
hype/hope for opportunistic device to device file sharing with rpi3.

On Thu, Jun 23, 2016 at 3:23 AM, Peter Robinson 
wrote:

> On Thu, Jun 23, 2016 at 10:36 AM, Adam Holt  wrote:
> > Congrats to the F24 Team:
> >
> > The strategic question from a broad deployment perspective across the
> > world's lower-middle class is increasingly Raspberry Pi support,
> according
> > to so many grassroots/field groups I'm speaking with in 2016, very
> > increasingly tempted to dump Fedora/CentOS for Raspbian, no matter what I
> > tell them, so that learning technologies are not impeded by traditional
> > education bureaucracy -- coming into schools thru the front/back and side
> > doors.
> >
> > I personally hope this "only Raspbian can save us" sentiment is
> premature,
> > in that I don't see the Raspbian ecosystem as being fully
> mature+resilient
> > just yet -- and as such I *hope* CentOS (or Fedora, or Debian, or...)
> > deliver increasingly competitive offerings on RPi 3, RPi 4, RPi 5 (or
> > similar) into 2020~
>
> I have most of the bits in place for the RPi2/3 in F-24, just ran out
> of time in the lead up to Beta to land the last bits. We will have
> F-24 images for them soon and OOTB support in F-25.
>
> P
>
> >
> > From: Matthew Miller 
> > Date: Tue, Jun 21, 2016 at 9:47 AM
> > Subject: Fedora 24 is here!
> > To: annou...@lists.fedoraproject.org
> >
> > Today the Fedora Project is pleased to announce the general release of
> > Fedora 24. Download it now from our Get Fedora site:
> >
> >  Workstation: https://getfedora.org/workstation/
> >  Server:  https://getfedora.org/server/
> >  Cloud:   https://getfedora.org/cloud/
> >
> >  Spins:   https://spins.fedoraproject.org/
> >  Labs:https://labs.fedoraproject.org/
> >  ARM: https://arm.fedoraproject.org/
> >
> >
> > Another Step in the Fedora Journey
> > --
> >
> > The Fedora Project has embarked on a great journey... redefining what
> > an operating system should be for users and developers. Such innovation
> > does not come overnight, and Fedora 24 is one big step on the road to
> > the next generation of Linux distributions. But that does not mean that
> > Fedora 24 is some "interim" release; there are great new features for
> > Fedora users to deploy in their production environments right now!
> >
> >
> > Workstation
> > ---
> >
> > The Fedora 24 Workstation release features GNOME 3.20, with many
> > usability improvements such as easier input device and printer
> > settings, a better search interface, shortcut windows for keyboard
> > commands, and more convenient music controls.
> >
> > Flatpak (formerly xdg-app) is another building-block feature, with
> > Software able to track installed Flatpaks and adding more features in
> > the future as the technology develops. The Software app has also grown
> > features to provide a full system upgrade directly from the desktop
> > from one Fedora release to the next, and the ability to provide
> > labeling as well as reviews of available software.
> >
> > Fedora 24 continues our work on the X replacement, Wayland, a
> > next-generation graphics stack. Although this release will not default
> > to Wayland, it includes many improvements and is available as an option
> > for users to try out, and potentially will be the default stack in
> > Fedora 25.
> >
> >
> > Server
> > --
> >
> > Fedora 24 Server edition is more streamlined and introduces more
> > modularity, which will become a major factor in future Fedora releases,
> > even as unnecessary packages were removed and the installer has a
> > smaller footprint.
> >
> > FreeIPA 4.3 is a major feature for Fedora 24 Server.
> > FreeIPA is an integrated security information management solution. This
> > new version of FreeIPA features simplified replica installation and
> > improved replication technology management.
> >
> >
> > Cloud
> > -
> >
> > Fedora is on its way to being the best platform for containerized
> > applications, from base Fedora container images to a full-featured
> > platform as a service to run and manage them.
> >
> > As we continue on this part of the journey, we are packaging OpenShift
> > Origin so it is easy to deploy. OpenShift Origin is a Platform as a
> > Service system based around Kubernetes, a production-grade container
> > orchestration project. OpenShift Origin is optimized for application
> > development and deployment. Origin makes it easy for developers to get
> > started building applications in containers and for operators to manage
> > them.
> >
> > While not shipped in Fedora 24, per se, we have new infrastructure for
> > developing container images with applications layered on top of the
> > base Fedora 

Re: [Server-devel] [XSCE] Re: [Sugar-devel] Sugar-Server enhancement

[George Hunt]

Problem 1:
Sugar now has a "register again" which I have only used to connect to the
same server. I guess that it's hard to do a ssh connection to a different
server (same IP) without ssh complaining about a possible man in the middle
attack. This is the situation described in
https://bugs.sugarlabs.org/ticket/362. Manash is proposing to fix that.

When I read the flow chart at (1) I don't know whether it will work. Port
5000 is not a jabber server (see 2). The Authserver at (2) was written by
activitycentral, I don't know gunicorn or flask, I have not been able to
understand how it works, and I've never successfully interfaced with it.
So it makes sense to try to use port 5000.  Whether it will provide a list
of connected clients? I'm not sure (jabber seems more appropriate).  Maybe
some on this list can suggest a easy jabber client/library (nothing hit me
when I googled "python jabber client")

I'd be willing to join in a larger effort to address the problem Tony is
pointing out.  But maybe we should let Manash fix bug 362 as a well defined
task before we tackle  problems 2 and 3 below.

Background as I see it:
Problem 2:
Contrary to original OLPC assumptions, some user situations have many
children per laptop.  Should our server interactions accommodate that need?
I think we should try. I think Tony is asking us to move past original
basic assumptions.

It sounds to me that, with James suggestion about using lightdm, (which I
also have played around with), and his knowledge that on SOAS, most of the
Activities still work (with username != olpc), we have most of a solution
to the multiple identities on a piece of hardware, connected to the single
server. I know it's wrong to jump too quickly to implementation, but we
could modify the pam module makehomdir to do its normal function, as well
as enable lightdm, and disable olpc-dm.

Problem 3:
A very small number of XO's are used in a school server setting I think.
Even so, any elastic boundary between local and server journal data store
may make sense. I've heard Tony's idea about needing a journal "keep" flag
for years. And I think it's a good one. But putting it on the details page,
and not changing all the documentation relating to "favorites" flag, also
makes sense. Once we have identities unique and defined (solve problem 1
and 2), I think Tony already has working code that solves problem 3.

(1)
https://wiki.sugarlabs.org/go/Features/Multiple_schoolserver_registration
(2) https://github.com/XSCE/xsce/blob/master/roles/authserver/tasks/main.yml


On Wed, Apr 13, 2016 at 5:30 PM, Adam Holt <h...@laptop.org> wrote:

> On Wed, Apr 13, 2016 at 8:17 PM, Tony Anderson <tony_ander...@usa.net>
> wrote:
>
>> James
>>
>> I can't think of a use case for an XO to access multiple school servers.
>
>
> Apologies am not following all the details of this conversation, but just
> to point out some of our newer work in Haiti will experiment with XO
> laptops moving between different school servers, in off-campus computer
> club(s), librar(ies), and at school, etc.
>
> Likewise there are definitely teachers who work in multiple schools, and
> need to bring their main XO to each school, impressive!
>
> Also some schools campuses are just too large to tie in all the IT infra
> together, with buildings just too far apart.  So there will be several
> inexpensive school servers likely in these schools, who do not want to
> network their buildings together at this point.
>
> Finally George Hunt is advising us on how to construct even more mobile
> servers too, based on a large USB battery packs supporting Raspberry Pi 3
> (or XO-4 or whatever) permitting increasingly ruggedized knowledge hotspots.
>
> *Not sure how we will deal with naming all these different school servers
> coming down the pike, but definitely something to start thinking about now,
> a great question :}*
>
>
>> If a school is so large (> 200 XOs), the easy solution is two school
>> servers. However, the XOs access the schoolserver based on SSID and so
>> users can be divided to their own schoolserver by the connection - defining
>> the LAN.
>>
>> The OLE Nepal solution clearly satisfies the requirement to avoid any
>> unnecessary configuration option or 'onboarding' step.
>>
>> The current placement of the code has been working for years. Showing the
>> server in the network section was added at a late release, as I recall
>> because the register menu item disappeared after registration and so a user
>> did not know for sure if the XO was registered. This was later fixed with
>> 'register again' so putting the name in the network section is now
>> redundant.
>>
>> Tony
>>
>> On 04/14/2016 07:35 AM, James Cameron wrote:
>>
>>> On Wed,

Re: [Server-devel] Alas, I'm giving up on Access Point function for XO1.5

[George Hunt]

The kernel which has thinfirm is at:
http://download.unleashkids.org/xsce/downloads/os/xo1.5/kernel-3.3.8_xo1.5-20160102.1250.06CE.8ddad46.i686.rpm

And there's an Internet In a Box image with AP function all set up at:

http://download.unleashkids.org/xsce/downloads/os/xo1.5/160210_xo1.5_ap.img.zip
ready to dd onto an SD card (minus content).

On Sun, Apr 3, 2016 at 4:39 PM, James Cameron <qu...@laptop.org> wrote:

> Thanks, very interesting.
>
> The wireless card contains an ARM processor, and that's what the
> firmware runs on.
>
> Yes, the firmware had proven to be stable when used by Open Firmware
> for NANDblaster function, but Open Firmware uses a subset of the
> commands supported.
>
> On the other hand, Linux uses many more commands, and so is likely to
> expose any undiagnosed problems in the firmware, which cause the card
> to hang and the driver to cease traffic.
>
> Some work was done upstream in the past year for mwifiex/mv8787
> recovery after firmware hang, but I've not noticed any recent work on
> libertas.
>
> My gut feel is that your random observation ping results are missed
> receive interrupts; a packet is being received late just after the
> system sends the card an unrelated command, and you might observe a
> response to tuning of the ping interval.
>
> Do you have your latest kernel source available anywhere?
>
> On Sat, Apr 02, 2016 at 11:44:03AM -0700, George Hunt wrote:
> > The recent SD cards are so cheap, and wikipedia in various languages is
> such a
> > good replacement for the encyclopedia I devoured as a kid, I spent a few
> weeks
> > on this project.
> >
> > The results were mixed, but just not good enough. The XO1.5 can act as a
> > gateway for a few minutes, or  a few hours (with apparently random
> delays to
> > the point of failure). But eventually, the libertas thinfirm driver, or
> > firmware, gets to be unresponsive, and only a reboot will correct the
> > situation.
> >
> > History
> > The libertas AP function was developed by OLPC in the 2010 timeframe
> (and used
> > successfully by openfirmware for flashing many XO at a time). I'm not
> clear
> > whether "thinfirm" firmware was generated by marvel, or modfied by
> outside
> > consultant, cozybit under nondisclosure agreement., This firmware may
> have
> > removed some function, but added AP capability. The driver was modified,
> in
> > order to add the sdio interface (the OLPC motherboard uses SDIO interface
> > rather than the pre-existing usb driver interface). The thinfirm driver
> (with
> > USB interface) has been upstreamed, and accepted into the kernel.  But
> the SDIO
> > interface modifications to the driver have not. More recently discovered
> sdio
> > code has not been tried (2) -- or maybe it's the same in a different
> package.
> >
> > Olpc documentation of the AP function existed, and I tried to follow the
> > instructions (1)(4)(7)
> >
> > Jerry Vonau found a patch for sdio interface to thinfirm, which was the
> one I
> > tested (5). If I were to continue, I'd probably try the patch at (2)
> since it's
> > more recent.
> >
> > There are references to setting the libertas_tf driver into debug mode,
> which I
> > did not pursue, since I'd need to learn a lot more about wireless
> protocols, to
> > make use of the debugging information.(6) But that's probably what I
> would need
> > to do if I continued.
> >
> > Random observation:
> >if a wifi client is connected to an XO1.5 acting as AP, and both the
> client
> > and the server are set to ping one another, the client ping delays are
> > constant. but the server delays are bi-modal, either 5ms or 230ms.
> >
> >  1. [1]http://wiki.laptop.org/go/Thinfirm_1.5
> >  2. [2]http://www.sleepyrobot.com/?p=160 --
> >  3. [3]
> http://dev.laptop.org/git/users/derosier/wireless-testing/tree/drivers/
> > net/wireless/libertas_tf
> >  4. [4]http://wiki.laptop.org/go/Libertas_Thinfirmware_HOWTO
> >  5. [5]http://www.spinics.net/lists/linux-wireless/msg55208.html
> >  6. [6]
> http://labs.isee.biz/index.php/How_to_setup_Marvell_88w8686_SDIO_wifi#
> > How_to_debug_the_libertas_driver
> >  7. [7]http://wiki.laptop.org/go/XO_as_AP
> >
> > References:
> >
> > [1] http://wiki.laptop.org/go/Thinfirm_1.5
> > [2] http://www.sleepyrobot.com/?p=160
> > [3]
> http://dev.laptop.org/git/users/derosier/wireless-testing/tree/drivers/net/wireless/libertas_tf
> > [4] http://wiki.laptop.org/go/Libertas_Thinfirmware_HOWTO
> > [5] http://www.spinics.net/lists/linux-wireless/msg55208.html
>

[Server-devel] Alas, I'm giving up on Access Point function for XO1.5

[George Hunt]

The recent SD cards are so cheap, and wikipedia in various languages is
such a good replacement for the encyclopedia I devoured as a kid, I spent a
few weeks on this project.

The results were mixed, but just not good enough. The XO1.5 can act as a
gateway for a few minutes, or  a few hours (with apparently random delays
to the point of failure). But eventually, the libertas thinfirm driver, or
firmware, gets to be unresponsive, and only a reboot will correct the
situation.

*History*
The libertas AP function was developed by OLPC in the 2010 timeframe (and
used successfully by openfirmware for flashing many XO at a time). I'm not
clear whether "thinfirm" firmware was generated by marvel, or modfied by
outside consultant, cozybit under nondisclosure agreement., This firmware
may have removed some function, but added AP capability. The driver was
modified, in order to add the sdio interface (the OLPC motherboard uses
SDIO interface rather than the pre-existing usb driver interface). The
thinfirm driver (with USB interface) has been upstreamed, and accepted into
the kernel.  But the SDIO interface modifications to the driver have not.
More recently discovered sdio code has not been tried (2) -- or maybe it's
the same in a different package.

Olpc documentation of the AP function existed, and I tried to follow the
instructions (1)(4)(7)

Jerry Vonau found a patch for sdio interface to thinfirm, which was the one
I tested (5). If I were to continue, I'd probably try the patch at (2)
since it's more recent.

There are references to setting the libertas_tf driver into debug mode,
which I did not pursue, since I'd need to learn a lot more about wireless
protocols, to make use of the debugging information.(6) But that's probably
what I would need to do if I continued.


*Random observation:*
   if a wifi client is connected to an XO1.5 acting as AP, and both the
client and the server are set to ping one another, the client ping delays
are constant. but the server delays are bi-modal, either 5ms or 230ms.

   1. http://wiki.laptop.org/go/Thinfirm_1.5
   2. http://www.sleepyrobot.com/?p=160 --
   3.
   
http://dev.laptop.org/git/users/derosier/wireless-testing/tree/drivers/net/wireless/libertas_tf
   4. http://wiki.laptop.org/go/Libertas_Thinfirmware_HOWTO
   5. http://www.spinics.net/lists/linux-wireless/msg55208.html
   6.
   
http://labs.isee.biz/index.php/How_to_setup_Marvell_88w8686_SDIO_wifi#How_to_debug_the_libertas_driver
   7. http://wiki.laptop.org/go/XO_as_AP
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] olpc-dm/Xorg segfault only if during systemd boot

[George Hunt]

I missed one obvious clue, thinking that Xorg.log was the strategic place
to look.

/var/log/messages (search olpc-dm) shows

   - "olpc-dm.service holdoff time over, scheduling restart"
   - and then "xauth: file /var/tmp/olpc-auth/.Xauthority does not exist"

googling "Xauthority create olpc" does not give a good hit, but suggests I
should learn olpc-utils better.

On Mon, Sep 7, 2015 at 2:44 PM, James Cameron  wrote:

> I'm interested in finding the cause; any such race condition that
> prevents X startup may eventually occur depending on timing, so is
> there some simple reproducer that we can apply to OLPC OS?
>
> Where are your changes relative to OLPC OS?
>
> What does the systemd journal say about olpc-dm service?
>
> How do the journal, olpc-dm logs, and X logs differ between working
> and failing case?
>
> Where does gdb show the segmentation fault occurs?
>
> --
> James Cameron
> http://quozl.linux.org.au/
>
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Re: Rpi2 now available, and loads current XSCE master

[George Hunt]

I recently attended the SF OLPC monthly meeting, and Alex Kleider was
asking for help getting the wifi going.

So before I offered to help, I came home and downloaded the image that I
had worked on back in March (almost forgotten). (just dd if=
of=>/dev/sdb or whatever> bs=4M)

When I just went to the docs on our git repo, I experienced the bad links.
The link to the available image is at
http://download.unleashkids.org/xsce/downloads/installer/rpi2/  and the
inadequate documentation is at
https://github.com/XSCE/xsce/blob/master/docs/INSTALL.rst in the middle of
the page.

On Sun, Sep 6, 2015 at 11:58 AM, Andreas Gros  wrote:

> Hi George,
> I'm trying to install XSCE on an RPi2 for our project in Ethiopia, but all
> the links I've found for images/packages on the wiki and in the post below
> don't work anymore.
> Is there an updated version/repo/build/link somewhere?
>
> Thanks a lot!
>
> Andi
>
>
> On Thursday, February 26, 2015 at 10:00:09 AM UTC-8, georgejhunt wrote:
>>
>> Just got my second rpi, and discovered that they both failed until I used
>> a more robust USB to micro USB power cable.
>>
>> It's pretty fast, for just 3watts, including the USB wifi dongle.
>>
>> To get the full install on fc21, I needed to comment out ajenti at
>> /opt/schoolserver/xsce/roles/8-tools/meta/main.yml
>>
>> The base software, with rpi loader and kernel, is available at
>> http://www.digitaldreamtime.co.uk/images/Fidora/21/ (based on a minumal
>> 173MB download)
>>
>> There's a rtc that mounts on the accessory connectors of the rpi at
>> http://www.amazon.com/gp/product/B00MCNMR8G
>>
>> I'm still fiddling to get a USB wifi driver, and there's still lots of
>> testing. But completing the install process is significant in itself. Once
>> I get it all tested, I'll post an image. I'll be about 2GB in size.
>>
>
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Access Point capability for the XO1.5 -- the dream of internet in a box on an SD card

[George Hunt]

Back in the 2010 timeframe, there was an effort to add Access Point
capability to first the XO1, and later the XO1.5. Of particular interest to
me is [1], which documents in detail how to install and configure the
Thinfirm libertas driver. I found evidence in the git repo [2] that the
driver was built and tested upon kernel version 2.6.22.

Javier configured the kernel via [3], which I think I have applied to
kernel 3.3.8 (the kernel used in FC18, and the most recent 13.2.5 releases).

I have attempted to follow the instructions in release notes [4]:

   - modprobe libertas_tf loads mac80311, and libertas_tf
   - put the firmware in /lib/firmware/sd8686tf.bin
   - But no interrupts show up in /proc/interrupts and no devices apparent
   to iwconfig.

Does anyone know that thinfirm does or does not function on kernel-3.3.8?

Is there someone who has more knowledge who would be willing to
coach/assist?


[1] http://wiki.laptop.org/go/Thinfirm_1.5
[2]
http://dev.laptop.org/git/users/javier/libertastf.git/tree/arch/x86/configs/i386_defconfig
[3]
http://dev.laptop.org/git/users/javier/libertastf.git/commit/?id=5d4134fa4b524995a73ec282b613a48709c5859c
[4] http://dev.laptop.org/pub/firmware/libertas/thinfirm/RELEASE_NOTES_SDIO
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Trying to access a school server from the outside world

[George Hunt]

I thought I had already set you up for using the Amazon passthrough. But
here are the steps:


   1. There is a user at the https://50.17.210.12:943/admin/ port with
   username:gonzalo and a password I will send separately.
   2. Sign on there and change your password. You can use this sign on at
   any time to see if the sora server client connection is available at the
   passthrough.
   3. Then you will need to create a new user without administrative
   privileges that will become the server's client connection to the vpn. The
   generation of the keys for the sora server is triggered by accessing
   https://50.17.210.12:943/ (without admin), and logging on with the
   credentials you created when you set up the sora server user.
   4. When you make this https:// access, the amazon openvpn application
   will offer to let you download the openvpn client application.  I have
   usually yum installed openvpn already. Hit refresh, and you will be given
   a choice to download an unattended access key file.
   5. Download the cient.ovpn file and change it so something similar to
   the username you created. Place it in the /etc/openvpn/ directory of sora
   server

Sorry I missed your request when it came 3 days ago.


On Sat, Jun 13, 2015 at 7:54 PM, Gonzalo Odiard godi...@sugarlabs.org
wrote:

 Could I use your passthrough server to access Sora server?
 What we should do setup it?

 Gonzalo

 On Sat, Jun 13, 2015 at 4:16 PM, George Hunt georgejh...@gmail.com
 wrote:

 Typically a server is behind some sort of NAT device, and some sort of
 firewall, and most likely has a variable ip address assigned by the ISP's
 dhcpd.  The trick is to have the server initiate an outgoing conversation
 to a device on the internet that is always on.  I purchased a micro
 instance on amazon cloud for the purpose.

 The amazon instance generates keys for clients which permits passthrough
 conversations between any clients. There's two levels of authentication --
 1. need a vpn key to connect to the amazon instance, and 2. need
 authentication at the ssh port of the target (preferably a public key in
 .ssh/authorized_keys on the target -making dictionary attacks less likely).

 But I'm becoming a fan of teamviewer. You need to install Xorg, and I
 usually install XFCE because it's pretty light weight. Up until now, I've
 resisted a GUI for servers.



 On Sat, Jun 13, 2015 at 2:25 PM, Tim Moody t...@timmoody.com wrote:

 I should also have mentioned that we have started using TeamViewer on
 some of the servers which allows a session on the server without using the
 vpn hub.

 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel



 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel




 --
 Gonzalo Odiard

 SugarLabs - Software for children learning

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] XO1.5 as Access Point using thinfirm

[George Hunt]

Thanks to James Cameron's work on FC22 kernel, I have an image of XSCE
running in an SD card on an XO1.5.

The regular libertas driver works as a client.

There is documentation for using a special driver, and special firmware at

http://wiki.laptop.org/go/Libertas_Thinfirmware_HOWTO:
http://wiki.laptop.org/go/Thinfirm_1.5
http://wiki.laptop.org/go/Test_Report

Which I have attempted to follow. I have added config items(below) and
created a new kernel.
Questions;

   1. The libertas_tf module loads via modprobe (after rmmod libertas
   cfg80211) Does this mean that it has found the firmware?
   2. After blacklisting libertas, and a reboot, there is a new device
   sit0, which may indicate that the proper network driver is not being
   found/loaded.(dmesg is quiet about libertastf)
   3. Are there incompatabilities between libertas_sdio and libertas_usb --
   should it be one or the other?


CONFIG_LIBERTAS_THINFIRM=m
CONFIG_LIBERTAS=m
CONFIG_LIBERTAS_USB=m
CONFIG_LIBERTAS_SDIO=m
CONFIG_LIBERTAS_DEBUG=y
CONFIG_LIBERTAS_MESH=y
CONFIG_FW_LOADER=y
CONFIG_MAC80211=m
CONFIG_MAC80211_HAS_RC=y
CONFIG_MAC80211_RC_MINSTREL=y
CONFIG_MAC80211_RC_MINSTREL_HT=y
CONFIG_MAC80211_RC_DEFAULT_MINSTREL=y
CONFIG_MAC80211_RC_DEFAULT=minstrel_ht
CONFIG_MAC80211_LEDS=y
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Two (minimal) goals for a kernel for XO1.5 based upon fc22?

[George Hunt]

If I set just the goals of getting keyboard input, and display output, what
problems will I face trying to use defconfig_xo1.5 from dev.laptop.org
(x86-3.3 branch)?

Does anyone have a config file that works on the XO1.5 on a Fedora release
later than FC18 that I might springboard from?

So far, I've just been training myself about kernels, required drivers in
the kernel, and dracut. It looks like most of the required code is
upstreamed in FC22. But the kernels I've built so far based upon FC22 have
not met either of the two goals I set.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] httrack web based interactive site crawler

[George Hunt]

Tim

Adam wanted me to discuss centos with Fiji. So I tried to do an install of
current master on Centos.

The kalite-serve enable timed out, and it looks from the log that it might
have been creating datasets when ansible yanked it away.  Have you
discovered how to recover from this situation?

On Thu, Apr 23, 2015 at 10:45 AM, Tim Moody t...@timmoody.com wrote:

 My plan has been to add this capability to the admin console.  I was
 thinking just recursive wget, but if there is something higher level that
 would be great.

 So far you can download and install zims and launch Kalite to get videos.
 I want to add the collections Rachel, IIAB, and Bernie (if Tony is
 interested ) and then move on to html-based sites.

 Sent from my BlackBerry® PlayBook™
 www.blackberry.com

 --
 *From:* Anish Mangal anis...@umich.edu
 *To:* xsce-devel xsce-de...@googlegroups.com, server-devel 
 server-devel@lists.laptop.org
 *Sent:* 23 April, 2015 7:07 AM
 *Subject:* [XSCE] httrack web based interactive site crawler

 Hi,

 I found this amazing software that can be configured and interactively
 crawls and keeps offline copies of websites.

 http://www.httrack.com/page/1/en/index.html

 Is useful for including custom content. An educator could bundle a website
 of their choice using this tool, which, though not perfect for cases where
 there are youtube videos, is useful for other media - pdf, documents etc.

 Perhaps there is a case for including this in xsce?

 Best,
 Anish


___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [UKids] Enthusiasm for a Centos School Server is wonderful

[George Hunt]

And testing is going to be a lot rough at first. I started using Tim's
suggestion to build up an automated server from
https://github.com/XSCE/xsce-tests. But when it didn't work, I discovered
that I was using an XO1.5, and the code had been written for an XO1, but
the documentation didn't indicate anything about the hardware platform.

In any case, if you clone https://github.com/XSCE/xsce-tests, you don't
need to go through all the process of making a stick, and loading it on a
XO1 using tiny core linux. The useful scripts are in the /testing folder.
You can use git directly on an XO to fetch them.

And then the scripts themselves will need to be modified I think, based
upon which server mode you are testing (appliance, lan-controller, or
gateway).  An appliance exists on an infrastructure which already has
gateway, dhcp, dns, etc). A lan-controller is sufficient to itself, and
gateway is a lan-controller plus. Obviously not all test are appropriate in
every mode.

I will be the guinea pig, and start testing my rpi2 installation, and share
with you all the trials and tribulations. For that purpose I have created a
public spread sheet, (now mostly empty) which I will use to document my
testing progress. I invite everyone to contribute and tweak that page, and
maybe collectively, we can make it a snapshot of what works and doesn't.

This new publicly writeable sheet (with this link) is at
https://docs.google.com/spreadsheets/d/18Y07kuAiHyMGJNQLydgo7GbuHShu_F5BQWA3egt_1PU/edit?usp=sharing

On Sat, Mar 28, 2015 at 10:29 AM, Tim Moody t...@timmoody.com wrote:

 and, of course, each application needs testing for functionality apart
 from did it install and does it run.



 *From:* Danishka Navin [mailto:danis...@gmail.com]
 *Sent:* Saturday, March 28, 2015 1:24 PM
 *To:* Tim Moody
 *Cc:* Adam Holt; unleashk...@googlegroups.com; xsce-devel; server-devel
 *Subject:* Re: [UKids] Enthusiasm for a Centos School Server is wonderful







 On Sun, Mar 29, 2015 at 1:11 AM, Tim Moody t...@timmoody.com wrote:

 At this time, and subject to testing, xsce can be installed on Fedora 18
 (for XOs), 20, and 21 and on Centos.  Any and all of these needs testing.



 Keep in mind that xsce is not necessarily bundled with one of these OS in
 that you can install the os and then runansible or one of the other setup
 scripts.



 However, to make life easier, George has images and an installer that you
 just stuff onto a machine and have a working xsce.  Similarly Jerry has an
 iso that already has all the needed rpms and other installables and will
 give a livecd style ready to go xsce.



 However, you also need to keep in mind that if you install from an image
 or an iso you are getting a particular os and a snapshot of xsce.



 If you want to roll your own, you can first install an OS.



 I use what's on XO-4 for F18, minimal for F20, and as Danishka mentions
 Server Edition for 21.  I haven't tried Centos.



 Finally, please also keep in mind that all of these things have happened
 since the last release (5.1) of xsce so are a work in progress, which is
 exactly why a lot of testing is needed.







 Tim,

 Do we have list items to be test? test case?



 --
 Unsung Heroes of OLPC, interviewed live @ http://unleashkids.org !
 ---
 You received this message because you are subscribed to the Google Groups
 Unleash Kids group.
 To unsubscribe from this group and stop receiving emails from it, send an
 email to unleashkids+unsubscr...@googlegroups.com.
 For more options, visit https://groups.google.com/d/optout.

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Rpi2 now available, and loads current XSCE master

[George Hunt]

Just got my second rpi, and discovered that they both failed until I used a
more robust USB to micro USB power cable.

It's pretty fast, for just 3watts, including the USB wifi dongle.

To get the full install on fc21, I needed to comment out ajenti at
/opt/schoolserver/xsce/roles/8-tools/meta/main.yml

The base software, with rpi loader and kernel, is available at
http://www.digitaldreamtime.co.uk/images/Fidora/21/ (based on a minumal
173MB download)

There's a rtc that mounts on the accessory connectors of the rpi at
http://www.amazon.com/gp/product/B00MCNMR8G

I'm still fiddling to get a USB wifi driver, and there's still lots of
testing. But completing the install process is significant in itself. Once
I get it all tested, I'll post an image. I'll be about 2GB in size.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] easy or complicated to change XSCE server + OpenVPN names?

[George Hunt]

Somehow I missed this email.  I agree with Tim, that we need to standardize
a place for a version number of the school server.

I can create openvpn keys for all three schoolservers, and we can preload
them into /etc/openvpn.  Then all that is required is to change the
variable in /usr/bin/xs-vpn, to point to the correct software key.


On Fri, Jun 27, 2014 at 9:13 AM, Tim Moody t...@timmoody.com wrote:

   the name of the schoolserver should never be changed, though its domain
 can be changed from .lan to something else.  so far we have not worried
 about multiple servers having the same name because they are not expected
 to be on the same network, but at some point we should probably introduce a
 subdomain, especially where more than one server belongs to the same
 organization.

 I leave it to George to rule on the openvpn user name for the server, but
 changing that seems possible.

 What I worry about is that we are currently on track to deploy two
 different versions of xsce this round.  At the very least we should store a
 version string somewhere and make sure the setup document for each server
 records this.

 Tim

  *From:* Adam Holt h...@laptop.org
 *Sent:* Thursday, June 26, 2014 7:56 PM
 *To:* server-devel server-devel@lists.laptop.org ; xsce-devel
 xsce-de...@googlegroups.com
 *Subject:* [XSCE] easy or complicated to change XSCE server + OpenVPN
 names?

  George,

 Just curious in case Sora/Haitians desire to re-order deployments,
 wondering how much hassle she/they'd face:

 EG. she's scheduled to deploy school delmas28 July 5+, school
 boisdavril July 12+, and school lascahobas a week later July 19+ ---
 but logistics can change as ~3 servers+all are quickly canoe'd to Haiti
 over the coming month ;)

 (Or if it's too complicated+dangerous to have her change server  OpenVPN
 names on the ground in Haiti, just let us know :)

 --
 Unsung Heroes of OLPC, interviewed live @ http://unleashkids.org !

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] NUC and solar power questions

[George Hunt]

On Fri, Jun 20, 2014 at 7:09 AM, Ben Burrell baburr...@gmail.com wrote:

 George:
 I have a physics student who will be assembling and testing a solar setup
 similar to that described on the Unleash Kids site (
 http://www.unleashkids.org/2014/02/21/off-the-grid-suitcase-solar-solutions-for-power-in-haiti/).
  I understand that the system can re-charge about 10 xo1's each day with 6
 hrs of sunlight where the battery is discharged no more than 50% (one of
 the things she'll test).
 Can you give me some estimates about what is needed to power the NUC
 you're testing, a hard drive with iiab, and an access point?


The average NUC power is about 9 watts, and the AP that I have running from
village telco draws 1.5 watt (and runs on 9-15 volts -- good for lead acid
batteries).  I don't think village telco units are easily available from
amazon.  Maybe someone else knows.

 1. Approx what # of watts per hour do the panel(s) need to generate to run
 the system for, say, 3 hours per day...assuming 6 hours of sunlight and
 maintaining a proper discharge of the battery.

Well, if you need 300 whr for  10 XO's, and 30jwhr to run the NUC for 3
hours, the bar minimum is 330 watt hours per day. Then you need to factor
in losses in charging of battery, overcast days, etc.  So if you had a 100w
solar panel working for 6 hours per day, with clouds, overhanging trees,
lack of proper sun orientation, you'd probably be lucky to get 330 watt
hours per day


 2. Is the 10 amp charge controller (described on the website) sufficient
 or would a 20 amp be better?


depends a lot on the size of your solar panel. In my fixed single 100 watt
solar installation, I've never seen more than 65 watts (5Amp).

 3. Is more than one 12V battery needed?

batteries come in different capacities.  If you want to handle a few cloudy
days the trojay 105 AHR 12v battery should really only be used to store 600
watthours (use half). With losses, that's not really enough for 2 days
overcast. So probably two 6 volt trojan t-105's, which are rated at 225AHr,
might be a better choice


 4. Can you power an access point from DC  (please pardon my ignorance)  If
 not, is there a low cost true sine wave inverter that you can recommend?


It's worth your time to find a 12 powered AP I think. I've been impressed
with the support we can get for village telco units.


 Thanks.
 Ben


 On Fri, Jun 20, 2014 at 6:10 AM, George Hunt georgejh...@gmail.com
 wrote:

 Tony,

 I'm currently running the i3 NUC off of a solar charged LA battery. Boots
 fine. Per Richard Smith's analysis, I'd say that the NUC can handle the 14+
 charging/equalizing cycles, but might not get a full 12v in really cold
 climates (with the internal 2.5 hard disk running on the 5 volt rail, and
 the motherboard probably using buck converters to much less than 12, I'd
 expect it to work well).



 On Fri, Jun 20, 2014 at 12:42 AM, Tony Anderson tony_ander...@usa.net
 wrote:

 George,

 Great news! It sounds like this may turn into the hardware of choice for
 a school server in the coming
 year. I'll try to see how to get one in EU - may have to wait until I
 get back to Germany in a few weeks.

 If possible, could you verify that it boots and operates from a 12v
 battery? This is the environment in the Kokobe school in Lesotho. I would
 like to swap out the MSI and operate a server directly from the battery.

 Tony


 On 06/20/2014 01:54 AM, George Hunt wrote:

 the i3 NUC arrived, has an upgraded bios, and is running the hard drive
 that I removed from the MSI, successfully headless. The annoying thing is
 that they changed the video output to micro hdmi.  Fortunately I had an
 adapter.

 the Documentation is specific that it will run on 12-19.  I was
 concerned that lead acid batteries might go lower than that during deep
 discharge, but referring to the charts suggests that at capacity in amp
 hours divided by 20 gets down to 12 volts at 30% remaining charge. Docs at
 http://downloadmirror.intel.com/23089/eng/D54250WYB_
 D34010WYB_TechProdSpec05.pdf

 The NUC idles at 8 watts, spiking to 16 when it's doing anything
 intense.



  --
 Unsung Heroes of OLPC, interviewed live @ http://unleashkids.org !
 ---
 You received this message because you are subscribed to the Google Groups
 Unleash Kids group.
 To unsubscribe from this group and stop receiving emails from it, send an
 email to unleashkids+unsubscr...@googlegroups.com.
 For more options, visit https://groups.google.com/d/optout.



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Server Backup on HaitiOS

[George Hunt]

Tim discovered that backups to the XSCE in release 5.0 were failing.  I had
changed ds-backup-server to use the WSGI interface (mod_python was
obsoleted in FC18). So I assumed  that the ball was in my court. But I
believe the test just completed indicates that the problem was really that
the superuser bit in /bin/ping was removed in FC17 (the base for HaitiOS).

The test:

   1. Load OS13.2.0 on an XO4.
   2. Install XSCE release 5.0 on top of that -- installs ds-backup-server
   hash c5d86 (unchanged from XSCE-0.4)
   3. Load HaitiOS hash 6d78 on an XO1
   4. Register XO1 to server
   5. Wait for 2 hours -- observe no data in directory at
   /library/user/SHC
   6. execute chmod 4755 /bin/ping on the XO1
   7. Wait for 1 hour -- observe that the backup had occurred

I think the wifi fixup that Tim has prepared for Sora to take to Haiti
should include the change in permissions on /bin/pin.

I do remember that, in the Tiny Core environment, it is hard to find the
file that I wanted to change (because I don't have a clear idea of the disk
tree, before the chroot that the firmware does as it is bringing up the XO.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [UKids] School Server Call: Nepal focus (Thurs 10AM NYC Time, June 5th)

[George Hunt]

Request for purchase information below:


On Thu, Jun 5, 2014 at 2:25 AM, Adam Holt h...@laptop.org wrote:

 Nepal will be our focus today, with core developer Sumit Acharya joining
 us live from Nepal working thru the evolution of OLE Nepal's many schools
 servers.

 With OLE Nepal's new MSI DC111 servers remaining cheap at just over $200
 but now having 1.8GHz, 500GB HDD, 2 GB RAM and USB3, we will talk about
 upgrading Fedora Commons and Fez compatibility as many worldwide are
 looking to upgrade to Fedora 18 if not Fedora 20 if not CentOS..


Amazon has MS DC111 for $350. Does anyone know how we can get them for
$200+ in the US?

 .

 Please join in less than 8 hours from now, adding Agenda items here!

 https://docs.google.com/document/d/1o6QtzLb6e58YKWqMf_junux2XyBRLFm31un8YLcYslg

 Please reply privately with your Skype or phone number -- backchannel is
 #schoolserver on http://webchat.freenode.net

 --
 Unsung Heroes of OLPC, interviewed live @ http://unleashkids.org !

 --
 Unsung Heroes of OLPC, interviewed live @ http://unleashkids.org !
 ---
 You received this message because you are subscribed to the Google Groups
 Unleash Kids group.
 To unsubscribe from this group and stop receiving emails from it, send an
 email to unleashkids+unsubscr...@googlegroups.com.
 For more options, visit https://groups.google.com/d/optout.

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] server and IAB

[George Hunt]

The fact that the screen comes up on a client XO suggests that the server
is configured correctly.

Build 852 is a very old build. I looked in sugarlabs.org activities browse,
to see if there were any newer Browse versions that work with sugar .84.
 It does not look to be the case.

I suggest loading 13.2.0 on one XO, and see if that works. see
http://wiki.laptop.org/go/Release_notes/13.2.0

Or in Haiti, we solidified around 12.1.0, because the move to gtk version 3
in 13.2.0 seemed to add overhead, and slow everything down (see
http://wiki.laptop.org/go/HaitiOS)


On Thu, May 1, 2014 at 5:31 PM, Martiny JeanBaptiste 
martiny.jeanbapti...@dc.edu wrote:

 Hi Adam,

 We cannot play any of the Khan Academy video of the IIAB from an XO. The
 screen comes up but it never plays. The videos play without any issues on
 the schoolserver. Does the XO need a special plugin installed. Specs on the
 XO being used Build: 852 Sugar: 0.84.16.



 Thank you,

 Martiny Jean Baptiste

 IT Support

 Dominican College of Blauvelt

 470 Western Highway

 Orangeburg, NY 10962

 Office: (845) 848-4050

 Mobile: (845) 222-4550

 Email: martiny.jeanbapti...@dc.edu



 *From:* Joanne Clemente
 *Sent:* Thursday, May 01, 2014 2:24 PM
 *To:* Martiny JeanBaptiste
 *Subject:* FW: server and IAB







 Joanne Clemente, Ed. D.

 Director of Online Learning

 Center for Adult Student Education (CASE):

 Faculty Trainer, Blackboard/Moodle Administrator, MapWorks

 Dominican College

 470 Western Highway

 Orangeburg, NY 10962

 845.848.4017

 [image: still-logo-smaller]



 *It is better to be good than to be right.*



 *From:* hotmail_e5f197d31ca4c...@live.com [
 mailto:hotmail_e5f197d31ca4c...@live.comhotmail_e5f197d31ca4c...@live.com]
 *On Behalf Of *Tim Moody
 *Sent:* Thursday, May 01, 2014 1:47 AM
 *To:* Joanne Clemente; Server-Devel@Lists. Org
 *Subject:* RE: server and IAB



 It sounds like the server is working, but we need more information about
 how the xos connect and are configured.



 First, from a browser on the xo does http://schoolserver.lan return the
 portal screen?



 If not, I suggest that Martiny have a look at
 https://github.com/XSCE/xsce/blob/master/docs/INSTALL.rst to understand
 what sorts of networking are supported.



 Also, we need to know what version of the OS is on the xos.
 --

 Date: Wed, 30 Apr 2014 19:31:16 -0400
 Subject: Fwd: server and IAB
 From: h...@unleashkids.org
 To: joanne.cleme...@dc.edu
 CC: curtathomp...@gmail.com; t...@timmoody.com; georgejh...@gmail.com

 Joanne,
 I am offline in Nepal and many of us are travelling this week. Plz have
 Martiny write the exact problem/configuration
 to public mailing list server-devel@lists.laptop.org
 Thanks!

 -- Forwarded message --
 From: Joanne Clemente joanne.cleme...@dc.edu
 Date: May 1, 2014 3:04 AM
 Subject: server and IAB
 To: h...@laptop.org h...@laptop.org

 Hi, Adam

 We were finally able to get the IAB to work with the schoolserver. We can
 view the videos on the server but cannot see anything from the XO1 or the
 XO1.5 from either the browser or the gnome side.

 Is there something we need to install on them?

 Thanks,

 Joanne



 Joanne Clemente, Ed. D.

 Director of Online Learning

 Center for Adult Student Education (CASE):

 Faculty Trainer, Blackboard/Moodle Administrator, MapWorks

 Dominican College

 470 Western Highway

 Orangeburg, NY 10962

 845.848.4017

 [image: still-logo-smaller]



 *It is better to be good than to be right.*



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] XSCE on Cubox

[George Hunt]

We're making small steps. Tim took Jon's kernel work, and added in a root
file system, and zipped a 4Gb image.  I picked that up, and installed XSCE
5.0 (mostly successfully), and rezipped, and posted that up at
http://download.unleashkids.org/xsce/downloads/.

There is the beginnings of documentation for an install at
https://github.com/georgejhunt/xsce/blob/cubox/docs/CUBOX.rst.

I'm aware that not very many Cubox are available (Solidrun will deliver
more in the middle of May).-- but the quad core 2 GB memory hardware is
somewhat snappier that an XO4. (it gets to a login prompt in 17 seconds
running on an SD card). I'm looking forward to running off of a sata hard
disk.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] mea culpa -- my recent rpms moved to testing in XSCE

[George Hunt]

I woke up this morning with a wider focus than I had yesterday.

Earlier, I had been concerned that XSCE 5.0 had been broken for over a
month, because our rpms are no longer available on activitycentral.com. I
fixed that problem, and pushed it to tag 5.0 on github a couple of days ago.

But then yesterday, I probably broke 5.0 again by inserting some untested
new rpms into http://download.unleashkids.org/repos/xsce/xs-extra. I was
focused on getting ansible to run to completion on cubox, as a longer term
ongoing project. I had forgotten about all the end users who might suffer
from my mistakes, as they were installing 5.0 on other platforms.

The unleashkids site is new, and until today did  not have a safe place to
put testing code. I created a new yum repo definition at
http://download.unleashkids.org/xsce/repos/xsce-testing.repo and a new repo
tree at http://download.unleashkids.org/xsce/repos/xsce/testing/

Any testing on cubox will need to wget the repo definition to
/etc/yum.repos.d.

I'll try, but probably fail, to do better  next time.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Fwd: [support-gang] Number of XO-1 per AP on XSCE - one data point

[George Hunt]

This number resonates with me.  Kevin Gordon, in Toronto, was interested in
XSCE primarily because it would offer ejabberd, without which the factorial
N conversations taken two at a time would fill the air waves at N=13.

I'm responding this way because of your statement that the clients were
non-XOs. My guess is that no AP will be able to handle more than 13
without ejabberd (and the registration process between the XO and the
server which enables it).


On Mon, Apr 7, 2014 at 12:14 PM, Adam Holt h...@laptop.org wrote:

 FYI!


 From: Nathan C. Riddle nathanr...@charter.net
 Date: Mon, Apr 7, 2014 at 12:05 PM
 Subject: [support-gang] Number of XO-1 per AP on XSCE - one data point
 To: Gang support-g...@laptop.org


 The maximum usable number of XO-1's (12.1.0) appears to be 14 using as AP
 TP-Link MR3020 on XSCE 0.4 RC1 on XO-1.5  with about 10 non-XO's showing in
 neighborhood. Number 15 is unreliable at connecting. With 14, simple
 connections to MOODLE appear reliable.   XSCE 0.4 was used since it was the
 only one available last September as school started.

 Testing DKMS libertas.ko file provided by quozl.org (
 http://dev.laptop.org/~quozl/12757/dkms with chmod 744 to file) to
 establish baseline for Mesh Potato-2 Basic AP replacement of MR3020.
 Presence of XO-1 with no modification appears to have no effect  on
 connections by modified XO-1's (as contrasted to XO-1's with old .ko file).

 Hoping MP-2 will push this past 25.

 Supplying this one data point since I had previously ask this question
 here.

 Nathan Riddle


 --
 Unsung Heroes of OLPC, interviewed live @ http://unleashkids.org !

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] cubox wifi

[George Hunt]

Jon,

In the long run, (my opinion), a cookbook description (similar to the
solidrun link) of the process would be more needed than a tarball.  Your
time, and our education, has real value.


On Tue, Feb 25, 2014 at 12:44 AM, Jon Nettleton jon.nettle...@gmail.comwrote:

 George,

 Things have actually changed a bit as the new kernel is using the new
 device-tree based configuration.  I will build you guys a testing
 kernel tarball and u-boot in a little bit.

 -Jon

 On Tue, Feb 25, 2014 at 12:18 AM, George Hunt georgejh...@gmail.com
 wrote:
  Hi Jon,
 
  Is it correct that I can follow the directions at:
 
 
 
 http://imx.solid-run.com/wiki/index.php?title=Building_the_kernel_and_u-boot_for_the_CuBox-i_and_the_HummingBoard
 
  with a clone from the github link you provided at
  https://github.com/linux4kix/linux-linaro-stable-mx6? Are there any
 gotchas?
  Recommendations?
 
  I forget the details from one month/day to the next.
 
 
  On Mon, Feb 24, 2014 at 9:38 AM, Jon Nettleton jon.nettle...@gmail.com
  wrote:
 
  Finally out and ready for use.
 
  https://github.com/linux4kix/linux-linaro-stable-mx6
 
  All the hardware is functional at this point.  The wifi driver is a
  bit chatty and needs a bit more love for AP mode.  Everything else
  from crypto to esata should be good to go.
 
  -Jon
 
  On Sun, Feb 23, 2014 at 5:28 PM, Jon Nettleton jon.nettle...@gmail.com
 
  wrote:
   It was probably a UHS capable card.  We found some hardware design
   problems that will need to be rectified to support UHS speeds for the
   sdcard.  For now we will be limiting the card-slot to standard SD high
   speed modes of 50Mhz which equates to around 20MB/s
  
   On Sun, Feb 23, 2014 at 5:12 PM, Tim Moody t...@timmoody.com wrote:
   sounds great.  one thing I noticed in Haiti was that I could not get
 a
   16G
   sd card to be recognized.  is this expected or was it my card?
  
   Tim
   -Original Message- From: Jon Nettleton
   Sent: Sunday, February 23, 2014 11:03 AM
   To: xsce-devel
   Cc: server-devel@lists.laptop.org
   Subject: Re: [XSCE] cubox wifi
  
  
   Tim,
  
   Actually the 3.10 kernel is done.  The wifi driver is a little chatty
   and I will work on sorting that out but it is functional.  I am just
   cleaning up the last few patches and will be pushing the kernel for
   initial testing in a few hours.
  
   -Jon
  
   On Sat, Feb 22, 2014 at 5:34 PM, Tim Moody t...@timmoody.com wrote:
  
   Jon,
  
   Can you (or anyone else) tell me where to find the drivers for the
   cubox
   wifi adapter you mentioned on the call on Tue.
  
   Thanks,
  
   Tim
  
  
 
 

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] cubox wifi

[George Hunt]

Hi Jon,

Is it correct that I can follow the directions at:


http://imx.solid-run.com/wiki/index.php?title=Building_the_kernel_and_u-boot_for_the_CuBox-i_and_the_HummingBoard


with a clone from the github link you provided at
https://github.com/linux4kix/linux-linaro-stable-mx6? Are there any
gotchas? Recommendations?

I forget the details from one month/day to the next.


On Mon, Feb 24, 2014 at 9:38 AM, Jon Nettleton jon.nettle...@gmail.comwrote:

 Finally out and ready for use.

 https://github.com/linux4kix/linux-linaro-stable-mx6

 All the hardware is functional at this point.  The wifi driver is a
 bit chatty and needs a bit more love for AP mode.  Everything else
 from crypto to esata should be good to go.

 -Jon

 On Sun, Feb 23, 2014 at 5:28 PM, Jon Nettleton jon.nettle...@gmail.com
 wrote:
  It was probably a UHS capable card.  We found some hardware design
  problems that will need to be rectified to support UHS speeds for the
  sdcard.  For now we will be limiting the card-slot to standard SD high
  speed modes of 50Mhz which equates to around 20MB/s
 
  On Sun, Feb 23, 2014 at 5:12 PM, Tim Moody t...@timmoody.com wrote:
  sounds great.  one thing I noticed in Haiti was that I could not get a
 16G
  sd card to be recognized.  is this expected or was it my card?
 
  Tim
  -Original Message- From: Jon Nettleton
  Sent: Sunday, February 23, 2014 11:03 AM
  To: xsce-devel
  Cc: server-devel@lists.laptop.org
  Subject: Re: [XSCE] cubox wifi
 
 
  Tim,
 
  Actually the 3.10 kernel is done.  The wifi driver is a little chatty
  and I will work on sorting that out but it is functional.  I am just
  cleaning up the last few patches and will be pushing the kernel for
  initial testing in a few hours.
 
  -Jon
 
  On Sat, Feb 22, 2014 at 5:34 PM, Tim Moody t...@timmoody.com wrote:
 
  Jon,
 
  Can you (or anyone else) tell me where to find the drivers for the
 cubox
  wifi adapter you mentioned on the call on Tue.
 
  Thanks,
 
  Tim
 
 

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] XSCE or XS7: Getting back to virgin state

[George Hunt]

I guess you'd need to look at the services individually, and determine what
you had changed during the testing, and reverse it.

I suppose you could run the ansible playbook again. It might  reinitialize
some files for you. Using the runansible script should not destroy the
function of the server. It should be able to run repeatedly. (though I saw
a  comment from Tim Moody yesterday that one of the services was erroring
out if ansible was rerun). (I have not tested this)

Of course you could reflash the XO, and start up the whole install process.
 But that would leave many things untested.

You mention XO registration.  You might try just running the server
registration tag: (I have not tested this)

ansible-playbook -i ansible_hosts xsce.yml --connection=local --tags=idmgr




On Sat, Feb 15, 2014 at 6:18 AM, tkk...@nurturingasia.com wrote:

 Need to some advice in schoolserver installation, testing and actual
 deployment.

 After XO registration on the school server.. and testing that all is well,
 what is the best way to restore the schoolserver back to its virgin state
 for the real deployment?

 Good weekend.

 ---
 T.K. Kang



 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] XSCE5: OpenVPN usage

[George Hunt]

Thanks Tim for your comment.  It jogged my memory, and I went back and
discovered a not printable character in one of the yml files which caused
runansible to error out. I pushed a corrected version to github.

Yes, the playbook I've posted connects to xsce.activitycentral.com.  The
amazon instance  that I set up for Haiti is one that I pay for ($130 per
year) to accommodate privacy concerns.


On Fri, Feb 14, 2014 at 9:07 AM, Tim Moody t...@timmoody.com wrote:

   I tried to install this when we got back from Haiti and ansible did not
 complete using runansible.  I also think it tries to connect to
 xsce.activitycentral.com rather than amazon.

 Tim

  *From:* George Hunt georgejh...@gmail.com
 *Sent:* Thursday, February 13, 2014 8:22 PM
 *To:* tkk...@nurturingasia.com
 *Cc:* Tim Moody timmo...@sympatico.ca ; Adam Holt h...@unleashkids.org
 *Subject:* Re: XSCE5: OpenVPN usage

  T.K.,

 Unfortunately, the openvpn feature, which was a part of 0.4 XSCE was not
 included in 5.0XSCE.

 When I was preparing for our Haiti deployment, I wrote up an Ansible
 playbook which included openvpn.  This is available at
 https://github.com/georgejhunt/ezadmin.  I think you can download the zip
 (available in the right hand column) file and install just the openvpn
 component by unzipping and running the following:

 ansible-playbook -i ansible_hosts ezadmin.yml --connection=local 
 --tags=openvpn

 Since openvpn was not part of the release cycle for XSCE 5.0, we may need
 to work together some to get it running.  I'm not sure the documentation
 applies to what I took to Haiti in every detail.

 On Thu, Feb 13, 2014 at 7:46 PM, tkk...@nurturingasia.com wrote:

 Dear George Hunt,

 Thanks for the information on the XSCE video chat regarding openvpn
 support if needed for the Malaysia deployment.

 Just started to read http://wiki.laptop.org/go/Setting_Up_an_XSCE_VPNtoget 
 started.

 On my XSCE 5when I type xs-vpn test as in the instruction I get command
 not found. Is VPN already installed with XSCE5 as I get the impression it
 is reading the instructions ? I try to add xs-vpn on but there is no
 existing  /etc/rc.d/rc.local file there.

 I am new to this and hence a little confused as to what is already there
 and if I am missing critical stps after a default XSCE 5 install.

 Cheers and some advice!





 ---
 T.K. Kang
 kk...@nurturingeducation.com





___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Re: a kernel for our solidrun protos?

[George Hunt]

Jon,

When I compiled the kernel that Rabeeh made available on github for the
Cubox, I discovered that the XSCE 0.5 ansible install errors out  because
iptables, and ajenti, do not install because of missing modules. In
attempting to modify the .config file (roughly 2000 lines long) I had very
slow and painful success.

I'm afraid that the window for addressing this sort of problem before we
leave for Haiti is closing.  Will we decide to purchase enclosures, hard
disks, etc, until we have some confidence that the XSCE software will run
on the Cubox's we have?

It seems clear to me that we will not  be able to count on having time to
work on kernel issues once we get to Haiti.

Can we all agree on a deadline, by which time we decide to go or no go for
using Cubox on this Haiti trip? My proposal is Jan 8. That will give us 10
days to order, and receive accessory disks, enclosures.

Jan 8 is 2 days from now!

What do others think?


On Fri, Jan 3, 2014 at 9:37 PM, Adam Holt h...@laptop.org wrote:

 Thanks Sameer, fyi George's XO-4 should arrive by Tues.


 On Fri, Jan 3, 2014 at 9:31 PM, Sameer Verma sve...@sfsu.edu wrote:

 On Fri, Jan 3, 2014 at 3:47 PM, George Hunt georgejh...@gmail.com
 wrote:
  On our skype call Thursday, Jon suggested using the XO4 config file as a
  base for getting all the needed kernel capability. And James suggested
  looking at the upstream kernel conf.  In any case I'll need to
 understand,
  and merge in, whatever special conig Rebeeh put in for the SOC in the
 Cubox.
 
  I won't have an XO4, until it arrives next week. (Adam agreed to send me
  one, if I'd build it up to XSCE for Haiti later this month). So here
 comes a
  request:
 
  Can I ask someone to email me the XO4 /boot/conf* file. I'm not at
 home, and
  I didn't bring an ARM XO with me to California. I would love to get
 XSCE to
  load, without erroring out, on a Cubox.
 

 George,
 If you are close by, you can borrow mine until Adam's XO-4 arrives.

 Sameer

  Do we need to start thinking about eSata interfaces. I happen to have an
  eSata enclosure from Startech, and a eSata cable.  But I didn't find any
  reference that looked like the kernel was recognizing an eSata
 interface. I
  searched the config I'm using. There were SATA config lines which
 were not
  enabled.
 
  Jon, what hardware have you been using? Maybe I can get you to email or
 post
  your config file.
 
 
 
 
  On Thu, Jan 2, 2014 at 12:16 AM, James Cameron qu...@laptop.org
 wrote:
 
  On Thu, Jan 02, 2014 at 12:00:43AM -0500, George Hunt wrote:
   Right now XSCE server software does not load, because some of the
   kernel networking userland (rtnetfilter) interfaces are not enabled.
 
  Speculation: the configuration changes between the upstream Linux
  kernel and the Fedora Linux kernel are what you may need to add.
  Looking at the kernel configuration (/boot/config*) of a Fedora kernel
  may give useful data.
 
  --
  James Cameron
 
 http://quozl.linux.org.au/

 --
 http://quozl.linux.org.au/
 Unsung Heroes of OLPC, interviewed live @ http://quozl.linux.org.au/
 http://unleashkids.org !


___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Cubox Solidrun Prototype Packaging

[George Hunt]

My experience tells me that connectors are the weak link in most systems.
 So I'm not excited about a connector that is part of the essential data
path, exiting the enclosure and exposed to being bumped/damaged.

Jon, I'd really like to start working on your kernel, and getting it to
play nicely with the XSCE. Is that possible? soon?


On Sat, Jan 4, 2014 at 10:14 AM, Jon Nettleton jon.nettle...@gmail.comwrote:

 I figured we just put the cubox at the back of the case for now and loop
 the esata cable back around into the box to hook up the drive.  Is there
 any reason to even worry about making the hdmi cable available?  In the
 back of my head I do have a dream where the XSCE could sit attached to a
 monitor/projector for teaching materials, but that seems far enough out not
 to worry about it.

 What does everyone else think?  You guys obviously have more in the
 field experience than myself.  Those sorts of cases just seemed like a
 relatively cheap solution to a problem that was brought up.

 I did talk with Rabeeh about the cooling for the i4pro and he said the
 external case didn't need any sort of special cooling it just radiates as
 is and should be fine completely enclosed.  There is actually a team that
 is building an i4pro into their RC surveillance plane :-)  My i4pro should
 be here this coming week and I can box it up, stress test it, and monitor
 thermals from the on-board sensor as well as my IR thermometer to make sure
 things look reasonable.



 On Sat, Jan 4, 2014 at 3:48 PM, George Hunt georgejh...@gmail.com wrote:

 What's the best strategy long term? The cubox has all its connectors on
 one face. Most of these need to be available to the outside. We can open
 the cubox and solder a esata connector that comes out the back. Picture is
 at http://imx.solid-run.com/wiki/index.php?title=CuBox-i_Hardware.

 Or we can try to project the essential stuff from the cubox mounted
 internally, and not really accessible from the outside.  The second item
 ending in 231 looks like it has pigtails to project the usb to the outside.
  We'd need to do the same for the internal ethernet adapter. This approach
 still leaves the hdmi, and console usb that might need to be available
 occasionally.

 So maybe in the long run, we need something intermediate between the two
 Jon has suggested, so that when it is necessary, we can get a hand into the
 box to make insert the hdmi or console cables.

  I'm in favor with going with number ..231 this time around.




 On Sat, Jan 4, 2014 at 8:56 AM, Jon Nettleton jon.nettle...@gmail.comwrote:

 Well from taking some basic measurements I was thinking we could pack a
 2.5 sata drive and connector along with the cubox into a 5.25 external
 case like this
 http://www.newegg.com/Product/Product.aspx?Item=N82E16817145090  That
 with a few dremel modifications and some soldering I think it would make a
 nice compact holder.  Given the time constraints that is probably not
 happening.  I did look around and found this which is more than enough
 space to do whatever is needed without much more cost,
 http://www.newegg.com/Product/Product.aspx?Item=N82E16811163231  It
 also has excellent ventilation.

 Of course these are just suggestions to help create a more
 self-contained product right now.




___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Cubox Solidrun Prototype Packaging

[George Hunt]

On Sat, Jan 4, 2014 at 11:49 AM, Jon Nettleton jon.nettle...@gmail.comwrote:

 As for the connectors, I don't think an esata connector is either less or
 more vulnerable than a cat5 connector, which we will have two of, one via
 usb, or any other connector.  Yes they are possible failure points, but as
 long as we make sure the esata cables have the locking connectors things
 should be fine.

 The new beta kernel is making progress.  I spent more time than I had
 wanted debugging some problems with another user that had a Cubox-i in
 regards to u-boot, device-tree, and the 3.10 kernel.  I have one last
 intermittent bug that I think is fixed and then it should be ready for
 testing.  If everything runs fine overnight it will be ready tomorrow
 morning.


That's excellent!




 -Jon


 On Sat, Jan 4, 2014 at 4:44 PM, George Hunt georgejh...@gmail.com wrote:

 My experience tells me that connectors are the weak link in most systems.
  So I'm not excited about a connector that is part of the essential data
 path, exiting the enclosure and exposed to being bumped/damaged.

 Jon, I'd really like to start working on your kernel, and getting it to
 play nicely with the XSCE. Is that possible? soon?


 On Sat, Jan 4, 2014 at 10:14 AM, Jon Nettleton 
 jon.nettle...@gmail.comwrote:

 I figured we just put the cubox at the back of the case for now and loop
 the esata cable back around into the box to hook up the drive.  Is there
 any reason to even worry about making the hdmi cable available?  In the
 back of my head I do have a dream where the XSCE could sit attached to a
 monitor/projector for teaching materials, but that seems far enough out not
 to worry about it.

 What does everyone else think?  You guys obviously have more in the
 field experience than myself.  Those sorts of cases just seemed like a
 relatively cheap solution to a problem that was brought up.

 I did talk with Rabeeh about the cooling for the i4pro and he said the
 external case didn't need any sort of special cooling it just radiates as
 is and should be fine completely enclosed.  There is actually a team that
 is building an i4pro into their RC surveillance plane :-)  My i4pro should
 be here this coming week and I can box it up, stress test it, and monitor
 thermals from the on-board sensor as well as my IR thermometer to make sure
 things look reasonable.



 On Sat, Jan 4, 2014 at 3:48 PM, George Hunt georgejh...@gmail.comwrote:

 What's the best strategy long term? The cubox has all its connectors on
 one face. Most of these need to be available to the outside. We can open
 the cubox and solder a esata connector that comes out the back. Picture is
 at http://imx.solid-run.com/wiki/index.php?title=CuBox-i_Hardware.

 Or we can try to project the essential stuff from the cubox mounted
 internally, and not really accessible from the outside.  The second item
 ending in 231 looks like it has pigtails to project the usb to the outside.
  We'd need to do the same for the internal ethernet adapter. This approach
 still leaves the hdmi, and console usb that might need to be available
 occasionally.

 So maybe in the long run, we need something intermediate between the
 two Jon has suggested, so that when it is necessary, we can get a hand into
 the box to make insert the hdmi or console cables.

  I'm in favor with going with number ..231 this time around.




 On Sat, Jan 4, 2014 at 8:56 AM, Jon Nettleton 
 jon.nettle...@gmail.comwrote:

 Well from taking some basic measurements I was thinking we could pack
 a 2.5 sata drive and connector along with the cubox into a 5.25 external
 case like this
 http://www.newegg.com/Product/Product.aspx?Item=N82E16817145090  That
 with a few dremel modifications and some soldering I think it would make a
 nice compact holder.  Given the time constraints that is probably not
 happening.  I did look around and found this which is more than enough
 space to do whatever is needed without much more cost,
 http://www.newegg.com/Product/Product.aspx?Item=N82E16811163231  It
 also has excellent ventilation.

 Of course these are just suggestions to help create a more
 self-contained product right now.






___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Re: a kernel for our solidrun protos?

[George Hunt]

On our skype call Thursday, Jon suggested using the XO4 config file as a
base for getting all the needed kernel capability. And James suggested
looking at the upstream kernel conf.  In any case I'll need to understand,
and merge in, whatever special conig Rebeeh put in for the SOC in the Cubox.

I won't have an XO4, until it arrives next week. (Adam agreed to send me
one, if I'd build it up to XSCE for Haiti later this month). So here comes
a request:

Can I ask someone to email me the XO4 /boot/conf* file. I'm not at home,
and I didn't bring an ARM XO with me to California. I would love to get
XSCE to load, without erroring out, on a Cubox.

Do we need to start thinking about eSata interfaces. I happen to have an
eSata enclosure from Startech, and a eSata cable.  But I didn't find any
reference that looked like the kernel was recognizing an eSata interface. I
searched the config I'm using. There were SATA config lines which were
not enabled.

Jon, what hardware have you been using? Maybe I can get you to email or
post your config file.




On Thu, Jan 2, 2014 at 12:16 AM, James Cameron qu...@laptop.org wrote:

 On Thu, Jan 02, 2014 at 12:00:43AM -0500, George Hunt wrote:
  Right now XSCE server software does not load, because some of the
  kernel networking userland (rtnetfilter) interfaces are not enabled.

 Speculation: the configuration changes between the upstream Linux
 kernel and the Fedora Linux kernel are what you may need to add.
 Looking at the kernel configuration (/boot/config*) of a Fedora kernel
 may give useful data.

 --
 James Cameron
 http://quozl.linux.org.au/

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Re: a kernel for our solidrun protos?

[George Hunt]

Thanks James



On Fri, Jan 3, 2014 at 7:06 PM, James Cameron qu...@laptop.org wrote:

 On Fri, Jan 03, 2014 at 06:47:25PM -0500, George Hunt wrote:
  Can I ask someone to email me the XO4 /boot/conf* file. I'm not at
  home, and I didn't bring an ARM XO with me to California. I would
  love to get XSCE to load, without erroring out, on a Cubox.

 Link rather than mail.


 http://dev.laptop.org/git/olpc-kernel/tree/arch/arm/configs/xo_4_defconfig?h=arm-3.5

 --
 James Cameron
 http://quozl.linux.org.au/

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Re: a kernel for our solidrun protos?

[George Hunt]

Jon,

Do you have a github url that includes the changes that you have made to
3.10.17 -- which I assume you lifted from
http://git.freescale.com/git/cgit.cgi/imx/linux-2.6-imx.git/tag/?id=rel_imx_3.10.17_1.0.0_beta
.

Now that Tim is scheduled to get his Cubox's, it seems like we should be
aiming as directly as we can for the goal.  I could create the github
presence, and you could contribute deltas.  Which direction is easiest for
you?


On Fri, Jan 3, 2014 at 7:23 PM, George Hunt georgejh...@gmail.com wrote:

 Thanks James



 On Fri, Jan 3, 2014 at 7:06 PM, James Cameron qu...@laptop.org wrote:

 On Fri, Jan 03, 2014 at 06:47:25PM -0500, George Hunt wrote:
  Can I ask someone to email me the XO4 /boot/conf* file. I'm not at
  home, and I didn't bring an ARM XO with me to California. I would
  love to get XSCE to load, without erroring out, on a Cubox.

 Link rather than mail.


 http://dev.laptop.org/git/olpc-kernel/tree/arch/arm/configs/xo_4_defconfig?h=arm-3.5

 --
 James Cameron
 http://quozl.linux.org.au/



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] XSCE 0.5

[George Hunt]

In order to change from a appliance machine to a gateway machine, the
playbook and ansible will need to rewrite the iptables, and get the ports
right so that dansguardian properly filters content that comes from the
internet.

I doubt very much that all this would work, without running ./runansible
again.  When I was loading 0.5 on a x86-64, the adapters were not
discovered properly.  The quickest way to discover this, for me, is to type
iptables-save in a root terminal.  Look at the masquerade line.  The output
-o needs to be the wan adapter, and the -i the lan.

I'm not understanding why you would want to change appliance - gateway
without running ansible?




On Mon, Dec 30, 2013 at 8:18 PM, Curt Thompson curtathomp...@gmail.comwrote:

  Just to recap for server-devel@lists.laptop.org,

 In virtualbox using XFCE and installing XSCE 0.5, a problem appeared with
 idmgr when doing ./runansible
 -the solution was to first run yum -y update audit and then rerun
 ./runansible.  After this, XSCE seemed to be working normally with the
 exception of the IIAB link being active despite no IIAB device attached.

 New Info:
 Setting the network adapter in Virtualbox to bridged made XSCE
 accessible from other devices on the LAN.  Rerunning ./runansible was not
 necessary (previously the adapter was in NATmode).  Appliance mode'
 seems to work quite smoothly and is easy to set up, in VirtualBox at least
 (will try on physical hardware later tonight).

 I had previously disabled SELinux, reenabled it to see if disabling it was
 necessary.  Many errors pop up when SELinux is enabled.  Set it back to
 disabled

 I enabled a second network adapter and booted up to see what would
 happen.  Everything seems to work locally, portal is still accessible from
 other machines on the host's network.

 Next question:
 However, XSCE does not seem to be acting as a server/gateway.  Would the
 best solution be to rerun ./runansible now that two network adapters are
 connected?  My understanding is that ./runansible will detect that there
 are two available adapters, one of which is an internet gateway, and set
 itself up in gateway mode.  Is that about right?  Would it be easy enough
 to do without rerunning ./runansible?  (Just get dhcpd going, for example?)

 ((I'm sort of imagining a real-world scenario where we got XSCE installed
 with one network adapter, then internet was cut off, then someone came
 through with a second USB network adapter - some scenario like that.  Could
 we get it going so that the machien would work as a router/gateway when
 internet came back... easily?))





 On 12/30/2013 4:29 PM, Anna wrote:

  As far as accessing it from other machines on your LAN, I've had success
 with configuring the VM's network for Bridged mode and specifying the host
 machine's adapter.  Can't remember if I needed to rerun ./runansible, but
 it seems like I didn't have to (though it wouldn't hurt anything).

  I also can't remember if I needed to disable SELinux on the host machine
 or not.  Sorry, I just got back home after a week with the family and my
 head's not quite back in the game yet.


 On Mon, Dec 30, 2013 at 6:20 PM, Adam Holt h...@laptop.org wrote:

 Curt,

 Can you share your conclusions more publicly on
 server-devel@lists.laptop.org or
 http://wiki.laptop.org/go/XS_Community_Edition or similar?


  On Mon, Dec 30, 2013 at 7:11 PM, Curt Thompson 
 curtathomp...@gmail.comwrote:

   Anna, George, Tim, Braddock,

 Thank you, ansible finished without issue after yum update audit,
 rebooted, and I seem to have XSCE 0.5 running in appliance mode - quite
 easy.

 URLs schoolserver and local IP 10.0.2.15 bring up portal as expected.

 semi-bug?
 Internet In A Box link appears active and is clickable though no IIAB
 device is attached.

 Next:
 What would I need to do to convert this into a Gateway-style install of
 XSCE?  I could enable another adapter in VirtualBox and rerun ansible, but
 the install documentation seems to hint that rerunning ansible may not be
 necessary.



 On 12/30/2013 3:28 PM, Anna wrote:
   Oh, you're right, Braddock, Miguel resolved that with pr/103:
 https://github.com/XSCE/xsce/pull/103
 
  So updating the audit package on FC18 should be the only workaround
 for a successful install, now.
 
 
   On Mon, Dec 30, 2013 at 5:22 PM, Braddock bradd...@braddock.com
 mailto:bradd...@braddock.com bradd...@braddock.com wrote:
 

  On 12/30/2013 03:20 PM, Anna wrote:
  I've used that ISO and for idmgr had to do:

  yum update audit

  Then there's a conflict with IIAB (which is currently being
  addressed, I believe) but in the meantime, the workaround is:

 What is the conflict, and am I the one addressing it?  I don't recall
 it off hand.

  -braddock

  On Mon, Dec 30, 2013 at 5:11 PM, Curt Thompson
   curtathomp...@gmail.com 
 mailto:curtathomp...@gmail.comcurtathomp...@gmail.com
 mailto:curtathomp...@gmail.com curtathomp...@gmail.com
 mailto:curtathomp...@gmail.com curtathomp...@gmail.com wrote:

 

Re: [Server-devel] a kernel for our solidrun protos?

[George Hunt]

After a few false starts:

   - Putting the microsd card upside down does not work.
   - The microusb console connector is really intermittent (does not seat
   well with my connector body).
   - My hdmi to vga adapter causes the monitor to jitter

I was not able to get the stock u-boot that Rabeeh provides to boot.
I downloaded the git repo, and selected the i4 make version.

Then for drill, I compiled his git repo kernel, and put it on partition 1.
I untared the rootfs onto a second partition.

At this point it's not clear that the kernel is able to get into the
rootfs, though the console messages say that it has discovered the mmcblk
device, and that it has mmcblk0p1, and mmcblk0p2.

I may need to change the kernel command line (I got to this point without
any modification of partition 1).

More play tomorrow



On Sun, Dec 29, 2013 at 10:45 AM, Jon Nettleton jon.nettle...@gmail.comwrote:

 George,

 Start out running the u-boot and kernel that Rabeeh has provided, that is
 based on the current stable kernel sources provided by Freescale.  I have
 a 3.10 based kernel that is more performant but I have not run it on the i4
 hardware yet.  Getting a baseline on the standard we can then better
 diagnose bugs or performance issues one way or the other with the new
 kernel.  For the second partition feel free to untar the F18 armhfp base
 image.  Changing the kernel command line can be done using a uEnv.txt file
 on the first partition of the SDHC port.  If we need anything more advanced
 I can write a u-boot boot script, or custom u-boot to handle loading
 things.  We may want a fallback kernel or something similar.

 Let me know if you have any questions or problems.

 Jon

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] a kernel for our solidrun protos?

[George Hunt]

Thanks for the pointers.  I did not know about uEnv.txt which I found at:

http://processors.wiki.ti.com/index.php/AM335x_U-Boot_User's_Guide#Setting_U-Boot_environment_using_uEnv.txt


On Sun, Dec 29, 2013 at 10:45 AM, Jon Nettleton jon.nettle...@gmail.comwrote:

 George,

 Start out running the u-boot and kernel that Rabeeh has provided, that is
 based on the current stable kernel sources provided by Freescale.  I have
 a 3.10 based kernel that is more performant but I have not run it on the i4
 hardware yet.  Getting a baseline on the standard we can then better
 diagnose bugs or performance issues one way or the other with the new
 kernel.  For the second partition feel free to untar the F18 armhfp base
 image.  Changing the kernel command line can be done using a uEnv.txt file
 on the first partition of the SDHC port.  If we need anything more advanced
 I can write a u-boot boot script, or custom u-boot to handle loading
 things.  We may want a fallback kernel or something similar.

 Let me know if you have any questions or problems.

 Jon

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] a kernel for our solidrun protos?

[George Hunt]

On Tue, Nov 26, 2013 at 4:09 AM, Jon Nettleton jon.nettle...@gmail.comwrote:

 I am working on that kernel a bit today so can whip up some images.  uboot
 needs to be dd'd onto the microsdhc card.  I will point you to the wiki and
 my binaries.

 -Jon


The Cubox is supposed to be arriving soon.  Is there anything that I can
start playing with? What is the status of your work?

George



 On Tue, Nov 26, 2013 at 2:23 AM, George Hunt georgejh...@gmail.comwrote:

 I've used xzimages, and zimages, on trimslice, and it was easy enough.
 Is uboot already on the board? Occasionally, I've needed to mess with
 uboot parameters, and I get lost very quickly in hex load addresses etc.
 Raspberry Pi starts off with a zimage and is trivial.

 But an image, and a short recipe would be good for me. I guess I was
 assuming that after getting the kernel, and a few basics, like yum, we
 could point to FC18 repos, and get to the base fedora 18 system, which we
 then augment with ansible to make it into a School Server.


 On Mon, Nov 25, 2013 at 1:15 PM, Jon Nettleton 
 jon.nettle...@gmail.comwrote:

 Do you guys need rpms or is a base zImage okay to get started?  I
 started getting together rpms but got buried in work.  Just compiling a
 zImage and .dtb file that you can drop in /boot would be easiest if you can
 work with that for starters.

 -Jon


 On Mon, Nov 25, 2013 at 7:00 PM, David Farning 
 dfarn...@activitycentral.com wrote:

 sani,

 What would you recommend?

 On Sun, Nov 24, 2013 at 7:43 AM, George Hunt georgejh...@gmail.com
 wrote:
  Jon,
 
  I'm beginning to want to do something with the solidrun proto. It
 looks like
  I can use the fedora 18 image at
 https://fedoraproject.org/wiki/Cubie_Board,
  and drop in a kernel that works with the new board.  Is that the
 case? Is
  there a kernel that I should be using?
 
  Or what do you suggest?
 
  George
 
  ___
  Server-devel mailing list
  Server-devel@lists.laptop.org
  http://lists.laptop.org/listinfo/server-devel
 



 --
 David Farning
 Activity Central: http://www.activitycentral.com





___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Re: $400 computer for Haiti

[George Hunt]

Looking at the pdf certification/specification documentation for the Dell
660s, I find the idle power is 51 watts, peaking to 100 watts under heavy
use.  The idle power continues to be my concern (Anna talks about heat. But
it amounts to the same concern). For comparison purposes, Tony's MSI Atom
bases machine measured 28 Watts, and the XO4 is about 8 Watts with external
hard disk. For the higher wattages, the fan becomes the weak link.

In my fantasy world, we could find the sweet spot between just the correct
amount of cpu for our server application, nestled into RAM and hard disk,
that minimized the cost of all the components.

But googling Zotac fan replacement, I came up with the deciding factor, for
me.  Zotac fans are not ball bearing, and apparently wear out within a year
time frame. They are not available from Zotac, and are not available in
replacement after-market form factor.

I have experience with Dell parts replacement. So I'm inclined to go with
Dell  the $399.99 Inspiron 660s i3-DualCore-3.4GHz-3MB cache, 4GB RAM, 1TB
7200rpm, 2X USB 3.0?
http://dell.com/us/p/inspiron-660s/pd

The Dell fan replacement instructions on page 59 of
ftp://ftp.dell.com/Manuals/all-products/esuprt_desktop/esuprt_inspiron_desktop/inspiron-660s_Owner's%20Manual_en-us.pdfmake
me much more confident that the hardware can live on for years.





On Thu, Dec 19, 2013 at 6:46 PM, Adam Holt h...@laptop.org wrote:

 What about the $399.99 Inspiron 660s i3-DualCore-3.4GHz-3MB cache, 4GB
 RAM, 1TB 7200rpm, 2X USB 3.0?
 http://dell.com/us/p/inspiron-660s/pd

 Or the $499.99 Inspiron 660s i5-QuadCore-3.3GHz-8MB cache, 8GB RAM, 1TB
 7200 rpm, 2X USB 3.0?

 http://dealnews.com/Dell-Inspiron-660-s-Ivy-Core-i5-Quad-2.8-GHz-Desktop-PC-for-500-free-shipping/935700.html
 http://www.logicbuy.com/deals/dell-inspiron-660s-deal/40218.aspx

 If we trust 7200rpm and Dell for reliability??  If so both appear to offer
 free shipping if ordered by Dec 20 11am EST.


 On Thu, Dec 19, 2013 at 6:38 PM, Anna ascho...@gmail.com wrote:

 What's the climate control situation at the site in Haiti?

 I live in a 100 year old house in Birmingham, Alabama, without central
 AC.  In the summer, I have to physically relocate equipment to my server
 room (a well insulated room where I keep a window unit on full blast).
 Otherwise, fans sound like jet planes taking off and everything gets really
 hot and stressed.

 Anyway, heat related failure would be one of my main concerns in Haiti.

 Anna


 On Thu, Dec 19, 2013 at 4:58 PM, George Hunt georgejh...@gmail.comwrote:

 Hi all,

 The Haiti deployment we are preparing for, made $400 available for a
 server, and power is available. What hardware has been the question.  Tony
 had lots of experience with Atom processors, and suggested nettop boxes. My
 research has led me to the following:

 $193 Zortec celeron processor
 http://www.amazon.com/Zotac-Dual-Core-Celeron-Barebone-ZBOXNANO-ID61-U/dp/B008OHRFE0/ref=sr_1_7?s=electronicsie=UTF8qid=1387490390sr=1-7keywords=zotac+barebone

 $75 8GB Memory --
 http://www.amazon.com/Corsair-1x8GB-Laptop-Memory-CMSO8GX3M1A1333C9/dp/B005T63BEM/ref=pd_sim_pc_1

 $86 WD blue 1TB drive --
 http://www.amazon.com/WD-Blue-Mobile-Hard-Drive/dp/B005DVJJWQ/ref=sr_1_4?ie=UTF8qid=1387491567sr=8-4keywords=wd+2.5+1+tb+blue

 Total $354 + tax

 Per Tim Moody's suggestion, I consulted
 http://www.cpubenchmark.net/cpu_list.php

 The Atom processor d535 has a passmark score of 693
 The celeron, for about the same price, selected in this proposal, has
 passmark score of 1215
 The i3 version of the Zortac box is $130 more and its passmark score is
 3833

 Of course, cpu benchmarks are not the only issue. I think it has been
 suggested more money should be put into memory, when trying to serve many
 clients.

 Soliciting inputs . . .

 George


 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel

 --
 http://lists.laptop.org/listinfo/server-devel
 Unsung Heroes of OLPC, interviewed live @
 http://lists.laptop.org/listinfo/server-develhttp://unleashkids.org !


___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] $400 computer for Haiti

[George Hunt]

Hi all,

The Haiti deployment we are preparing for, made $400 available for a
server, and power is available. What hardware has been the question.  Tony
had lots of experience with Atom processors, and suggested nettop boxes. My
research has led me to the following:

$193 Zortec celeron processor
http://www.amazon.com/Zotac-Dual-Core-Celeron-Barebone-ZBOXNANO-ID61-U/dp/B008OHRFE0/ref=sr_1_7?s=electronicsie=UTF8qid=1387490390sr=1-7keywords=zotac+barebone

$75 8GB Memory --
http://www.amazon.com/Corsair-1x8GB-Laptop-Memory-CMSO8GX3M1A1333C9/dp/B005T63BEM/ref=pd_sim_pc_1

$86 WD blue 1TB drive --
http://www.amazon.com/WD-Blue-Mobile-Hard-Drive/dp/B005DVJJWQ/ref=sr_1_4?ie=UTF8qid=1387491567sr=8-4keywords=wd+2.5+1+tb+blue

Total $354 + tax

Per Tim Moody's suggestion, I consulted
http://www.cpubenchmark.net/cpu_list.php

The Atom processor d535 has a passmark score of 693
The celeron, for about the same price, selected in this proposal, has
passmark score of 1215
The i3 version of the Zortac box is $130 more and its passmark score is 3833

Of course, cpu benchmarks are not the only issue. I think it has been
suggested more money should be put into memory, when trying to serve many
clients.

Soliciting inputs . . .

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] specs for the MSI unit Nepal is installing as schoolserver

[George Hunt]

Hi James,

Can you send me a link to the hardware that is being used in Nepal?  What
I'm currently looking at is
http://www.amazon.com/ZOTAC-Intel-Barebone-Mini-PC-ZBOX-ID41-U/dp/B004WO8O9Y/ref=zg_bs_3015426011_3
.

I thought 4 GB memory and a 1 TB drive might be needed.  But we have about
$400 budgeted and someone measured the power draw as about 25 watt.

Not as efficient as ARM, but ARM isn't really ready for prime time yet,
unless the cubox materializes quickly enough.

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Favorite School Server hardware?

[George Hunt]

Adam Holt is planning for an XSCE schoolserver deployment in Haiti in
January, and in this instance there's plenty of power, and my favorite
trimslice ARM may be under powered for the number of clients he is
contemplating.

The on-the-ground experience is not good with the trimslice that we
installed earlier this year in a Haiti deployment.  Reports are that no
more than 11 XO1 clients can simultaneously access the Internet In A Box.
We're not sure whether the bottleneck is at the wifi level, (XO's not
registered), the disk drive access, or the raw computing power of the dual
core ARM processor.

What other hardware have people been using recently?  What performance
measures does anyone have?

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Does XSCE need a new Home?

[George Hunt]

Adam, and I, talked for hours, primarily about the Malaysia summit, and
also a little about the hosting issue nicely explored in this thread.
 Thanks everyone.

What sticks in my mind is the idea that wiki.laptop.org is really strategic
for Miami, as long as they are in the business of selling XO's. There's
just so much hardware/software wisdom therein, and doubly so since many of
the people who had that information in their brains have gone on. (And
James Cameron has been doing a lot of tidying up -- I appreciate that).

I am struct by one of James Cameron's comments in this thread which I'd
like to pursue by branching off into another discussion.  His comment was

(I don't yet see how XSCE can benefit a tablet deployment, nobody
appears to have enumerated that.)

George



On Sun, Dec 1, 2013 at 5:41 PM, James Cameron qu...@laptop.org wrote:

 On Sat, Nov 30, 2013 at 12:44:46PM -0500, Tim Moody wrote:
  I'm in favor of staying where we are.  To me the biggest risk is
  that the site one day just disappears.  We should have a backup of
  the content and deal with it if and when it happens.  I'm not sure
  what the easiest way to do a backup is.

 Attached is a script that I use to update a local git repository with
 the wiki syntax form of the Forth related Wiki pages.  It might be
 extended to other pages.  Note the strange handling for multiple
 levels.  It might only handle one level, and might need further coding
 to fit your use case.

 --
 James Cameron
 http://quozl.linux.org.au/

 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel


___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Can XSCE benefit a tablet deployment?

[George Hunt]

This is a branch off of the thread Does XSCE need a new home?, and
stimulated by James Cameron's comments quoted where  in part:

I also disagree with the implication that the tablet is any more
proprietary than the laptop.  The mix of intellectual property is
certainly different, but the opportunity for use is also different.
Therefore the products shouldn't be compared at an intellectual
property licensing level.

What is very different is that the laptop was ground-breaking
original technology, and the tablet is buy-in technology.  This
cascades into completely different community involvement models.

(I don't yet see how XSCE can benefit a tablet deployment, nobody
appears to have enumerated that.)


At the SF summit, last month, Rodrigo demonstrated sugar running on a
google Nexus tablet. This was achieved by running the python/sugar
mega-package on top of the Ubuntu touch distribution.

Is the Nexus the right hardware platform long term? Well, it's
bootloaderis unlocked! And here, in my opinion, is where the licensing
becomes an
issue.

A little Vivitar digression:

The XO tablet is rebranded Vivitar. On Amazon, besides the XO Tablet, there
is also a lower cost/capability tablet introduced as a Camelo.  I
purchased one. Looked at it's End User Licensing Agreement. Under the
DCMA(Digital Millennium Copyright Act), it is illegal to reverse
engineer or
repurpose hardware that has proprietary material, without the express
consent of the manufacturer.  I contacted the Vivitar customer support for
the Camelo, and asked that they tell me how to unlock the boot loader --
that I liked their proprietary material, but I preferred to load other
software which was not covered by their license. They refused.

Whether Miami sees any value in giving its permission to unlock, and
repurpose the XO Tablet seems to me an open question. My own preference
would be to help increase manufacturing volumes of the XO tablet, and
continue to explore classroom technology integration of client and server,
which has really  only been successful in a few deployments.

At the Malaysia summit, there were a few Taiwanese tablet entrepreneurs,
pushing to become a hardware base for next generation classroom technology.
Whether any of these initial contacts could provide a way forward is my
current question.

I believe a tablet should be thought of as part of a learning technology
system, which also includes a laptop, and a school server. The tablet is
more appropriate for younger grades, and I believe will always a more cost
effective solution. Touch keyboarding, an essential skill for upward
mobility, can come with the addition of a bluetooth keyboard, or in upper
grades, a laptop. The school server is essential for both of these client
interfaces to manage media storage, limit pornography, collect usage
statistics, etc.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Does XSCE need a new Home?

[George Hunt]

Thanks Samuel,

Adam is coming through NYC today, and I'll spend some time with him.  I'll
try to get his take on the issues you discuss.


On Fri, Nov 29, 2013 at 6:44 PM, Samuel Greenfeld sam...@greenfeld.orgwrote:

 The following are my personal views, and not that of my previous or
 current employers.


1. At least for the near-term, all laptop.org hosting seems to be in
stasis.  It is not likely to go away, but without additional help it is not
likely to be kept up-to-date and/or improved.

My personal stance is that the items on paid hosting are at a higher
risk of suddenly disappearing than the items hosted by MIT (
dev.laptop.org, mailing lists, git, etc.).  Should OLPC-A decide to go
to a new primary webhost then the Wiki, RT, and a few other things could
disappear quickly and possibly without notice.

2. Given the lack of publicly known resources being provided by OLPC
to maintain Linux software development (besides low-level bugfixing), all
XSCE may be currently getting by using laptop.org is free hosting and
a spot under the domain name.

It may be worthwhile to have someone contact the business team at OLPC
to see what their long-term plans are for Sugar, XS, and open source
software in general.  From the outside it looks like almost all of their
resources have been shifted to the proprietary tablet product line.

XSCE may still wish to associate with OLPC.  But as best as I can tell
the corporate side of OLPC has already given XSCE and Sugar the cold
shoulder.

3. Personally I would prefer to see all non-commercial hosting
consolidated with Sugar Labs.  They already have a multi-site primary
infrastructure, download mirrors, and a group of system administrators who
regularly keep things up to date and monitor things.

Most importantly, almost all of their hosting is done by third parties
who do not have a financial interest in Sugar or XSCE.  I do not mind if
OLPC and/or Activity Central want to help host, but the community as a
whole should not be dependent on either being present.

4. We are never going to find hosting that is guaranteed never to go
away.  Witness Google's regular discontinuation of products, and
SourceForge's apparent decline in favor of GitHub.

It therefore may be prudent not to rely on any single host to be able
to provide a role for XSCE, and take an approach similar to Sugar Labs'
dual primary sites.






 On Fri, Nov 29, 2013 at 5:14 PM, George Hunt georgejh...@gmail.comwrote:

 Organizationally, I think there's a lot more leverage staying within the
 laptop.org fold, as long as it still exists. We don't need to repeat
 disconnects, and bad feelings, when there are really no egos involved. We
 have few enough people and resources that any dilution is a bad strategy.

 I think I over reacted to Samuel Greenfeld's concern about laptop.orghosting 
 going away. It seems to me that it's in everyone's interest to plan
 for an orderly transition from one hosting arangement to another, if that
 ever becomes necessary.  If during the interim, there needs to be bridge
 funding, I'd be glad to spearhead taking up a collection among the people
 that follow these things.

 George



 On Fri, Nov 29, 2013 at 4:06 PM, David Farning 
 dfarn...@activitycentral.com wrote:

 On Fri, Nov 29, 2013 at 7:13 AM, George Hunt georgejh...@gmail.com
 wrote:
  This is a more related to Samuel Greenfeld's comments than the XSCE
  documentation/install thread from which I branch.
 
  I think it makes sense to grab off of laptop.org all the essential
 stuff and
  put it some new place. We don't know the future, but in the present the
  prospect of ongoing support and maintenance of wiki.laptop.org seems
 more
  shaky to me.
 
  When I was first exploring the process of building something on top of
  schoolserver 0.7, I found ancestry of different components (documented
 at
  http://schoolserver.wordpress.com/xs-installation/rpm-heritage/). I
 think we
  should contact the most recent contributors to these ancestral repos,
 and
  with their cooperation/approval, move them to github.com.
 
  How to pay for a server somewhere? Three candidates come to me.
 Sugarlabs,
  activitycentral, and unleash kids.  Of these, activitycentral is the
 only
  for profit organization, I'd vote for the synergy that needs to exist
  between a for profit open source shop, and the ecosystem which it
 needs to
  remain legitimately part off a larger  open source community.


 The question in my mind is what would be best for XSCE. While I am
 personally investing in XSCE. I am also financially invested as XSCE
 is the community upstream to Dextrose Server. The better and more
 popular XSCE becomes, the more opportunities there are for AC to sell
 services.

 There seem to be three general options of XSCE. Stick with OLPC, go it
 alone, or partner with someone else in the ecosystem.

 So far

[Server-devel] Does XSCE need a new Home?

[George Hunt]

This is a more related to Samuel Greenfeld's comments than the XSCE
documentation/install thread from which I branch.

I think it makes sense to grab off of laptop.org all the essential stuff
and put it some new place. We don't know the future, but in the present the
prospect of ongoing support and maintenance of wiki.laptop.org seems more
shaky to me.

When I was first exploring the process of building something on top of
schoolserver 0.7, I found ancestry of different components (documented at
http://schoolserver.wordpress.com/xs-installation/rpm-heritage/). I think
we should contact the most recent contributors to these ancestral repos,
and with their cooperation/approval, move them to github.com.

How to pay for a server somewhere? Three candidates come to me. Sugarlabs,
activitycentral, and unleash kids.  Of these, activitycentral is the only
for profit organization, I'd vote for the synergy that needs to exist
between a for profit open source shop, and the ecosystem which it needs to
remain legitimately part off a larger  open source community.

George


On Fri, Nov 29, 2013 at 5:10 AM, Samuel Greenfeld sam...@greenfeld.orgwrote:

 I think you need to be careful how you phrase that -- you just half
 implied that all laptop.org hosting is going away.  There has been a fair
 amount of fear that resources may suddenly disappear, and I have been
 concerned about fragmentation where hosting of resources ends up all over
 the place.

 If there is a perceived need to migrate resources then that should be made
 clear, as others have already offered potential alternative hosting.  But
 there needs to be coordination.


 Focusing on the Deployment side, I would tend to agree with John's
 comments as well.

 Both Sugar and the Schoolserver have been historically focused on being
 their own ecosystem.  This has never changed, yet Sugar and the XS are
 often offered for use where existing DHCP, DNS, and other services already
 exist.

 Due to local policies, you may not be allowed to name your schoolserver
 schoolserver.  You may have to support 802.11x network authentication,
 etc.  It is possible to kludge these but the solutions are not elegant.

 If the Sugar and XSCE communities feel that the enterprise/first-world
 use case is a desired scenario where Sugar, IIAB, and/or Moodle may only be
 a part of a school's network instead of the primary role, then this
 specifically needs to be targeted.

 ---
 SJG



 On Fri, Nov 29, 2013 at 3:33 AM, James Cameron qu...@laptop.org wrote:

 I agree with John.  Every point in his documentation section should be
 handled.  Especially the point about wiki.laptop.org, which has so
 many distracting links on the navigation bar that we are all used to,
 but which new people become lost in.

 With regard to forums, the type that Google Groups has where they can
 also be received in mail may suffice.

 On Fri, Nov 29, 2013 at 11:44:33AM +0530, Anish Mangal wrote:
  Hi,
 
  I would like to share this blog post from John Ellis with the XSCE
 community:
  John is a high school student who is trying to setup XSCE in his
 class/school
  under the supervision of his teacher Jeff Elkner.
 
  http://johnmichaelffs.blogspot.in/2013/11/problems-with-xsce.html
 
  Some of the stuff he points out certainly makes a lot of sense to me, I
 think
  the core underling message is to make XSCE more approachable to the end
 user
  and the advanced end-user/deployer. He has gone to some lengths to
 point out
  specific aspects which could be improved.
 
  As we think about the possibilities for XSCE-0.6, I would like to
 further the
  discussion along these topics here and/or on IRC. I think the project
 could do
  well listening to end users' needs for the 0.6 cycle, especially that
 we now
  seem to have our house in order codewise thanks to the terrific work by
 all the
  software hackers here :-)
 
  Thoughts?
 
  Cheers,
  Anish
 
 
 

  ___
  Server-devel mailing list
  Server-devel@lists.laptop.org
  http://lists.laptop.org/listinfo/server-devel


 --
 James Cameron
 http://quozl.linux.org.au/
 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel



 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel


___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Does XSCE need a new Home?

[George Hunt]

Organizationally, I think there's a lot more leverage staying within the
laptop.org fold, as long as it still exists. We don't need to repeat
disconnects, and bad feelings, when there are really no egos involved. We
have few enough people and resources that any dilution is a bad strategy.

I think I over reacted to Samuel Greenfeld's concern about
laptop.orghosting going away. It seems to me that it's in everyone's
interest to plan
for an orderly transition from one hosting arangement to another, if that
ever becomes necessary.  If during the interim, there needs to be bridge
funding, I'd be glad to spearhead taking up a collection among the people
that follow these things.

George



On Fri, Nov 29, 2013 at 4:06 PM, David Farning dfarn...@activitycentral.com
 wrote:

 On Fri, Nov 29, 2013 at 7:13 AM, George Hunt georgejh...@gmail.com
 wrote:
  This is a more related to Samuel Greenfeld's comments than the XSCE
  documentation/install thread from which I branch.
 
  I think it makes sense to grab off of laptop.org all the essential
 stuff and
  put it some new place. We don't know the future, but in the present the
  prospect of ongoing support and maintenance of wiki.laptop.org seems
 more
  shaky to me.
 
  When I was first exploring the process of building something on top of
  schoolserver 0.7, I found ancestry of different components (documented at
  http://schoolserver.wordpress.com/xs-installation/rpm-heritage/). I
 think we
  should contact the most recent contributors to these ancestral repos, and
  with their cooperation/approval, move them to github.com.
 
  How to pay for a server somewhere? Three candidates come to me.
 Sugarlabs,
  activitycentral, and unleash kids.  Of these, activitycentral is the only
  for profit organization, I'd vote for the synergy that needs to exist
  between a for profit open source shop, and the ecosystem which it needs
 to
  remain legitimately part off a larger  open source community.


 The question in my mind is what would be best for XSCE. While I am
 personally investing in XSCE. I am also financially invested as XSCE
 is the community upstream to Dextrose Server. The better and more
 popular XSCE becomes, the more opportunities there are for AC to sell
 services.

 There seem to be three general options of XSCE. Stick with OLPC, go it
 alone, or partner with someone else in the ecosystem.

 So far we have chose to stick with OLPC. The benefits of building on
 the OLPC brand have been greater than the costs. The benefit is that
 the laptop.org site is still a 'hub' of the ecosystem. The cost is
 XSCE appears to be an unapproved and undesirable server clone piggy
 backing on the Association's message and brand. The question becomes,
 will the Association endorse XSCE as a successor or upstream to
 OLPC-XS or will it continue to live in the bowels of the wiki.

 Going it alone. This would be pretty straight forward. Getting a cheap
 VM somewhere or piggybacking on a larger project is possiable.
 Although it can be a lot of work and mean jumping though hops.  The
 questions becomes, do XSCE have the ability to thrive on it's own.
 There are about a dozen similar projects floating around in various
 states of completaion.

 Finally, there is partnering with someone with aligned goals such as
 Sugar Labs, unleash kids, or Activity Central.

 Whenever I think of the relationship between XSCE and AC, I think of a
 conversation I had years ago with Greg DeKoenigsberg (
 https://fedoraproject.org/wiki/Gdk/Experience ) about the relationship
 between Fedora and Redhat. He considered Fedora's relationship to
 Redhat it's greatest strength and its greatest weakness. The financial
 support and developer resources were valuable... but there were
 strings attached.

 We would be happy to host XSCE site on the AC infrastructure. We could
 handle it the same way we do the XSCE build VM.
 http://xsce.activitycentral.com/ is an independent VM with root access
 granted to community Sysadmins. DavidR, our web and communications guy
 could assist in setting up and migrating the site.


  George
 
 
  On Fri, Nov 29, 2013 at 5:10 AM, Samuel Greenfeld sam...@greenfeld.org
  wrote:
 
  I think you need to be careful how you phrase that -- you just half
  implied that all laptop.org hosting is going away.  There has been a
 fair
  amount of fear that resources may suddenly disappear, and I have been
  concerned about fragmentation where hosting of resources ends up all
 over
  the place.
 
  If there is a perceived need to migrate resources then that should be
 made
  clear, as others have already offered potential alternative hosting.
  But
  there needs to be coordination.
 
 
  Focusing on the Deployment side, I would tend to agree with John's
  comments as well.
 
  Both Sugar and the Schoolserver have been historically focused on being
  their own ecosystem.  This has never changed, yet Sugar and the XS are
 often
  offered for use where existing DHCP, DNS, and other services

Re: [Server-devel] a kernel for our solidrun protos?

[George Hunt]

I've used xzimages, and zimages, on trimslice, and it was easy enough. Is
uboot already on the board? Occasionally, I've needed to mess with
ubootparameters, and I get lost very quickly in hex load addresses
etc.
Raspberry Pi starts off with a zimage and is trivial.

But an image, and a short recipe would be good for me. I guess I was
assuming that after getting the kernel, and a few basics, like yum, we
could point to FC18 repos, and get to the base fedora 18 system, which we
then augment with ansible to make it into a School Server.


On Mon, Nov 25, 2013 at 1:15 PM, Jon Nettleton jon.nettle...@gmail.comwrote:

 Do you guys need rpms or is a base zImage okay to get started?  I started
 getting together rpms but got buried in work.  Just compiling a zImage and
 .dtb file that you can drop in /boot would be easiest if you can work with
 that for starters.

 -Jon


 On Mon, Nov 25, 2013 at 7:00 PM, David Farning 
 dfarn...@activitycentral.com wrote:

 sani,

 What would you recommend?

 On Sun, Nov 24, 2013 at 7:43 AM, George Hunt georgejh...@gmail.com
 wrote:
  Jon,
 
  I'm beginning to want to do something with the solidrun proto. It looks
 like
  I can use the fedora 18 image at
 https://fedoraproject.org/wiki/Cubie_Board,
  and drop in a kernel that works with the new board.  Is that the case?
 Is
  there a kernel that I should be using?
 
  Or what do you suggest?
 
  George
 
  ___
  Server-devel mailing list
  Server-devel@lists.laptop.org
  http://lists.laptop.org/listinfo/server-devel
 



 --
 David Farning
 Activity Central: http://www.activitycentral.com



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] xsce httpd.conf

[George Hunt]

Tim -- regarding httpd-xs.conf

My memory is that the issue httpd-xs.conf was really trying to address was
to set the memory limits, based upon the mount of total memory available --
getting squid, ejabberd, httpd, to share the available memory in an
equitable fashion. This seems more like a conf.d type of issue. And I also
remember how the roll over from httpd 2.2 to 2.4 caused us upgrade pain
that was really self inflicted. I don't think it's a good strategy to take
over the base conf file.


On Mon, Nov 25, 2013 at 3:01 PM, Tim Moody t...@timmoody.com wrote:

   PREAMBLE

 xsce has an httpd-xs.conf file in /etc/httpd/conf, which is a customized
 version of httpd.conf.  It gets used because /etc/sysconfig/httpd has a
 clause OPTIONS= -f conf/httpd-xs.conf.

 Except that in the ansible install /etc/sysconfig/httpd doesn’t get set.

 httpd-xs.conf is definitely needed, because it has a lot of settings not
 in the stock httpd.conf.  But because it replaces httpd.conf it also has
 lots of stock settings that are not unique to xsce.

 NOW THEREFORE

 We either need to modify ansible to set /etc/sysconfig/httpd or we need to
 put httpd-xs.conf into /etc/httpd/conf.d and have in contain only the
 settings we care about for xsce.  In the first approach we have basically
 taken ownership of all settings for httpd.  In the latter we have only
 taken ownership of the settings that are different in xsce, but we could
 get problems where settings accumulate rather than override (such as
 listen).

 So what’s the answer?

 btw while we are at it we should clean up the httpd-xs, xs-httpd, xs.conf,
 etc. proliferation of conf files.

 Tim

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] a kernel for our solidrun protos?

[George Hunt]

Jon,

I'm beginning to want to do something with the solidrun proto. It looks
like I can use the fedora 18 image at
https://fedoraproject.org/wiki/Cubie_Board, and drop in a kernel that works
with the new board.  Is that the case? Is there a kernel that I should be
using?

Or what do you suggest?

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Moving toward headless install on trimslice for XSCE

[George Hunt]

I'm wanting to do the following:

   - Get rid of the autologon to root at console tty.
   - Add a non privileged user (not a sudoer, not wheel), as we had
   pre-ansible, (user:admin,pw:12admin), so that the sshd config of
   permitrootlogon no can remain in place.
   - enable password authentication
   - Configure avahi to announce, so that in a trimslice situation, we can
   easily determine the remote sshd target ip.

What do people think?
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Ansible facts listing

[George Hunt]

I've been perplexed as to why the documentation about listing ansible facts
seemed to fail.

ie ansible localhost -m setup --connection=local returns no hosts found

What I discovered is that /etc/ansible/hosts file needs to be initialized
with:

[localhost]
127.0.0.1

Then the facts listing returns the desired information.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] extension of Moodle authentication mechanism

[George Hunt]

Thank you.  I'll spend some time studying it and try to write up some
documentation, and use cases.


On Mon, Oct 28, 2013 at 5:07 AM, Miguel González 
migonzal...@activitycentral.com wrote:



 On Mon, Oct 28, 2013 at 4:54 AM, George Hunt georgejh...@gmail.comwrote:

 I heard that someone at activitycentral was extending, augmenting the
 authentication used by Moodle, so that other web based services can climb
 on.


 It's called xs-autherserve and is a component in DXS initiative. If you
 install XSCE using ansible you will find it in
 http://schoolserver.local:5000.



 Can someone point me to the code?


 The source code is in https://github.com/migonzalvar/xs-authserver.



 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel




 --
 Miguel González
 Activity Central: http://www.activitycentral.com

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] what I'm working on

[George Hunt]

Hi xsce team,

I've been having fun working on little things that were not replicated by
ansible playbooks.  I started generating pull requests so everyone could
see what I'm working on, even though I'm leaving for a week in  Baja, and
the work is not really complete

But our new workflow really requires more testing and completeness than I
have achieved before I leave. I'm not generating pr's on incomplete work.

So here are the things I've got branches for, and done some work on:

   - openvpn
   - vnc-remote graphical administration
   - avahi-announce
   - a spec file that generates rpms that include hash tag, version,
   selectable base name
   - man-pages
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] aliases for hacking on our XSCE school server

[George Hunt]

I recently pushed an xs profile to github that will start fine tuning the
working environment on the XSCE Now I'm thinking that I should add and
collect some functions and aliases which make working with ansible
playbooks and github, easier, and require fewer keystrokes.

So far I've thought of:
runal = runansible $1 -- runansible-playbook with verbose diagnostic ouput
to $1 logfile.
gc $1 $2= git clone github.com/$1/xsce.git --depth=10 $2 -- for quickly
getting our own testing branch down to a unit under test
pullum = git checkout master;git pull upstream master ( might need to test
whether the first command succeeds before executing the second)
pushob = git push origin $1(branch)

The following aliases are currently available - though not collected into
one place in the profile
sr = systemctl restart $1
st = systemctl status $1
vlm = less /var/log/messages
du1 = du --max-depth=1 .
psg = ps -e|grep $1

I'm willing to collect everyone's ideas, test them, and update the system
profile.

Send me your suggestions.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] How to create a screencast

[George Hunt]

Santi,

In the demo last week, I think you said that you had used gstreamer to
generate screencasts, and that icecast might be used at the school server
end to distribute them, (there was some discussion whether icecast could do
multicast).

Can you give me a script, or at least more informtion about the gst-launch,
or other technique, that you used?

Thanks
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] extension of Moodle authentication mechanism

[George Hunt]

I heard that someone at activitycentral was extending, augmenting the
authentication used by Moodle, so that other web based services can climb
on.

Can someone point me to the code?
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Re: [IIAB] SGVHAK and IIAB Meeting this Saturday afternoon 10/26

[George Hunt]

Curt,

Checking the facts: You have a laptop, (what make and model), which has a
ethernet port, and a wifi adapter builtin. You want the wifi to act as AP.

If this is so, it's very similar to the situation we have with the
trimslice.  One issue is that the wifi chipset needs to be able to go into
access point mode. And the driver needs to be able to switch the hardware
into the AP mode.

The autoconfigure logic we wrote for XO's will not necessarily do the
correct things. So you will need to tell the network software how to
configure itselfr On the trimslice, I wound up needing to write
/etc/sysconfig/network-scripts/ifcfg-eth0 and ifcfg-wlan0 files.

see
https://github.com/XSCE/xsce/blob/master/plugins.d/startup/files/bootstrap-ts
-- the here-documents around line 60 show what worked for me. In
addition
you will have to load and run hostapd daemon  to handle connections that
other wifi clients will be trying to make.

George


On Fri, Oct 25, 2013 at 8:11 AM, Adam Holt h...@unleashkids.org wrote:

 On Fri, Oct 25, 2013 at 4:03 AM, Curt Thompson curtathomp...@gmail.comwrote:

 Hi Braddock.  Thanks for meeting with us.  You had asked if we were
 local in another thread -- I live right here in Pasadena and James is at
 LMU near LAX.

 This email mentions projects... I have Fedora-based XSCE School Server
 working with an Internet-in-a-Box hard drive, with an old wifi router as
 AP.  I have been trying to set up this laptop as an all-in-one
 server/router/wireless AP, where the the ethernet port would act as WAN
 with that connection shared via wireless (So, eth0 = WAN and wlan0=LAN,
 I believe).  Perhaps someone there would be willing to help me out with
 this.


 I'm asking the veterans at xsce-de...@googlegroups.com who might have
 more insight above?

 In any case, I look forward to meeting you and the rest of the group in
 a couple of days.

 Cheers,
 Curt

 On 10/24/2013 11:22 AM, Braddock Gaskill wrote:
  I will be at the SGVHAK meeting from approximately 3pm to 5pm this
 
  Saturday at the Hastings Branch Library and we can have an
 
  Internet-in-a-Box sub-meeting then.
 
 
 
  There will be two volunteers joining us from the One Laptop Per Child
 
  community who are working on IIAB deployments in Haiti and elsewhere.
 
 
 
  -braddock
 
  443-854-4691
 
 
 
 
 
  On Wed, 23 Oct 2013 21:58:42 -0700 (PDT), Lan Dang l.d...@ymail.com
 
  wrote:
 
  Reminder that we have a HAK meeting this Saturday at Hastings Branch
  Library.  We meet on the 4th Saturday afternoon of the month at the
  library.  (See end of e-mail for the specific dates.)
  I suggest bringing power strips and extension cords.  I've lost track
 of
  my giant power strip.
  When: Saturday, 10/26/2013 2:00 - 6:00 PM
  Where: Hastings Branch Library (3325 East Orange Grove Boulevard
  Pasadena,
 
  CA 91107)
  Come find us at the meeting room near the circulation desk.
  These are hands-on working meetings. There are tables, free wifi, and a
  group of knowledgeable people. So come, bring your projects, show off
  your
 
  new toys, and just connect with your fellow hardware geeks. There is
  even a
 
  projector if you want to do a mini-presentation.
  If you're planning on coming and working on something cool, please
 speak
  up.  We talked about trying to come up with a group project to work on
  during these meetings.  I think we could also help people finish up
  their
 
  projects.  It would give us a chance to learn from each other.
  It can sometimes be confusing to remember which Saturday we meet.   Our
  reservation goes from July 2013 to June 2014.  These are the dates I
  reserved:
  Oct 26, 2013 (this week)
  Nov 23, 2013
  Dec 28, 2013
  Jan 25, 2014
  Feb 22, 2014
  Mar 22, 2014
  Apr 26, 2014
  May 24, 2014
  Jun 28, 2014
  Lan





 --
 Unsung Heroes of OLPC, interviewed live @ http://unleashkids.org !

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] github merge of DXS and XSCE (including history of both) DONE

[George Hunt]

Hi folks,

After a small amount of head scratching (3hrs):

https://github.com/XSCE

is  set up, with history from both projects, as a base for all to
contribute to.

Thank you Jerry
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] github merge of DXS and XSCE (including history of both) DONE

[George Hunt]

Ok, Jerry just did the last step.  Neither of us was sure about the
conclusion of the long discussion we had about the issue.


On Thu, Oct 24, 2013 at 3:59 PM, Anish Mangal an...@activitycentral.comwrote:

 Not really. The xsce:dxs branch still needs to be merged *into*
 xsce:master afaict, so people can resume work on master (PR's and all).




 On Thu, Oct 24, 2013 at 12:15 PM, George Hunt georgejh...@gmail.comwrote:

 Hi folks,

 After a small amount of head scratching (3hrs):

 https://github.com/XSCE

 is  set up, with history from both projects, as a base for all to
 contribute to.

 Thank you Jerry




___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] github workflow

[George Hunt]

After some discussion at the sprint, I looked for documentation of the
workflow as I understand it:

https://www.atlassian.com/git/workflows#!workflow-forking

I found it helpful
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Re: XSCE Proposal | Moving to github and notifications

[George Hunt]

I like github, based upon my limited experience, because of its
browsability. There will be new workflows that need to be established, but
I'm willing for it.

My sense of it is that this becoming a activitycentral project, which is ok
with me. I'm glad when people start putting a shoulder to the wheel.

We need to start  adding value in the classroom as soon as possible, and
github is an intermediate objective, but a distraction, from that objective.

+1 for github

george


On Sun, Oct 13, 2013 at 2:11 PM, Anish Mangal an...@activitycentral.comwrote:

 *If* there is consensus on moving to github, should be think about making
 the switch before the hacksprint in SF? It might save us some time as then
 we would have set-up the buildbot, etc.

 On the other hand, if there isn't consensus, maybe the hacksprint would be
 a good opportunity to try github out.


 On Wed, Oct 9, 2013 at 8:30 AM, Martin Dluhos mar...@gnu.org wrote:

 On 10/09/2013 11:13 AM, Anish Mangal wrote:
  I think that it would be more appropriate to send notifications of
 merged pull
   requests rather than individual commits to avoid too much noise on
 the channel.
   Those who are interested can always follow the pull request link to
 access
   individual commits.
  
  I would prefer notifications for pull requests *both* when they are
  created and when they are merged. The creation of the pull request
  will sort of open a review and test window for people willing to do
  that, so notifying that event is necessary IMO.

 Very good point. I agree.
 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE]

[George Hunt]

 and holds the promise of incremental configuration.


 The XS install is a normal OS install in which a minimal Linux OS
 executes a script from a ramdisk to access the USB flash drive, decompress
 files and to copy them to the newly created partitions.

 Apparently, XSCE is installed by the previously installed OS (13.2). What
 is the dependency of XSCE on this OS? Does the XSCE install replace it?


 In the rpm and ansible scenarios it does not replace the OS on the XO or
 any other device.


 Naturally, the XO software build is for a client machine. It is based on
 graphical user interface, keyboard, and track pad.

 A server is intended to supply services to other computers (clients) via
 a network. As such, it is intended to be operated headless. Control of the
 server is normally by a remote login and command line operations.

 How does XSCE handle the change from a client machine to a server? Does
 it require a monitor and keyboard (or touchscreen)?


 Like XS, XSCE requires a monitor and keyboard (built into the XO) for
 installation, but can be basically headless thereafter.  Most of us turn on
 sshd and do the entire install remotely.  Probably other devices can come
 up with sshd capability already on and be entirely headless, if you can
 figure out their ip address.


 There is discussion of a home page. In XS-0.7 there is no home page. The
 only way to interact with the server is via ssh. Is this home page in XSCE
 the portal to a web site supported by XSCE or an administrative interface
 to XSCE?


 My experience of XS is that the home page is moodle.  I have been working
 on a high level home page aimed at end users from which there would be
 links to installed services.  It is intended to support local languages.



 For historical reasons, server administration is done in English (similar
 to the fact that air traffic control is done in English).
 I assume from this that a multi-language home page is the home page of a
 web site on the server, not server administration.


 yes.  however, several web-based administration tools are in the works.



 There has been much discussion of Ansible (vs Puppet). Is Ansible
 intended as a means to implement the build process for XSCE or as a means
 to implement and customize the install of XSCE or as a means to administer
 an XSCE installation?


 The implementer of ansible cleverly added an rpm build mechanism, but it
 is really aimed at installation, and, I hope, eventually at administration,
 either remote or local.


 The general philosophy of XS is to build a single installable version of
 the software (currently XS-0.7). After installation, scripts are
 provided to configure the software, primarily to match the local network.

 I think the assumption is that the server has ample disk capacity so that
 there is no need to limit the installed software packages to those that the
 deployment actually uses. For example, Moodle is installed but rarely used.
 XS-0.7 sets up a root partition of 8GB and the rest of the available
 capacity is given to the library partion.

 It is certainly understandable that this assumption would not apply to a
 school server based on an XO laptop.

 However, it has always been clear that the XO is a terrible choice for a
 server platform to be used only as a last resort. The primary reason for a
 school server is to give access to content beyond the limited capacity of
 the XO.

 Even though the OLPC project is intended to address the problem of the
 'digital divide', many developers seem to assume that deployed XOs have the
 same access to the internet as they themselves have. Even if a school has a
 DSL line shared by 100 XOs, that is not the same as a developer with access
 to a broadband network shared among a handful of users.

 I think it is important that the XSCE project define its goals and
 assumptions clearly.

 In the case of the XS-0.7, the assumptions are that the deployment has
 limited or no access to the internet, that it has a school server with
 large disk capacity, the server can support a local wifi network, the
 server can operate without system administration (essentially unattended
 except for power on and power off) for a full school year, the server can
 share internet access when connected to a WAN (ISP) with the XOs connected
 on the local network, and that the software on the school server be stable
 and updated at most once between school sessions.


 I think the XSCE community assumes that all of the above represents the
 most common scenario.  I grant that the requirement for large disk
 capacity, which I know you take to be axiomatic, is not met by all of our
 target platforms in ways that we find satisfactory yet.



 Tony


 On 10/11/2013 09:58 AM, George Hunt wrote:

 for Haiti, I just disabled multiviews in
 /etc/httpd/conf.d/xs-portal.**conf (commented out)


 On Fri, Oct 11, 2013 at 9:14 AM, Tim Moody t...@timmoody.com
 mailto:t...@timmoody.com wrote:

 just before I left

Re: [Server-devel] [XSCE] Dealing with huge xsce repo size

[George Hunt]

There's a political reason for the size of the repo.

In my discussions with Daniel Drake at the last SF summit, we discussed the
relationship between the work of XSCE and his finished 0.7 school server.
 The XSCE repo started from a clone of 0.7, which itself appears to trace
back all the way to the beginning of Martin's work on the server.

Daniel asked me to pretty up the XSCE commits, and collect them into
functional chunks, and submit them to Martin for review, as he had done
going from 0.6 to 0.7.  When I tried to learn, and use the git rebase
command, my lack of skill, and patience, came in the way of that objective.
 The number merge conflicts, and the need for merge-by-hand, seemed to me
to be just too likely to introduce errors, and the need for additional
debugging cycles.

At this point, the code base has diverged so much, I'm not willing to
rework the history as we discussed almost a year ago.

The reason to go slow in trimming down the size of the repo, from my point
of view, is that I'm not sure XSCE has replicated all the essential
functions of 0.7, particularly in the area of activation, lease management,
and anti-theft. These are essential features for large deployments.  Until
we learn about, and learn to test, in this area, we might want to keep
around the original functioning code.

George


On Sun, Sep 8, 2013 at 12:37 AM, Anish Mangal an...@activitycentral.comwrote:

 Hi,

 This is perhaps a very contentious topic, so I want to discuss with
 extreme caution :-)

 The size of the git repository for xsce is  70MB
 The actual size of the files is  3MB

 Now, I don't want to hurt anyone's sensibilities AT ALL here, but I feel
 70MB is quite a huge size for a repo containing code worth only 3 MB. I
 also feel it's a hindrance to keep code development nimble, making it
 unnecessarily difficult for users to download large repos. (Remember, if
 you're cloning a git repo, and you lose connectivity, you have to start
 over).

 Are there any thoughts we could improve the situation? Should it be
 improved?

 Best,
 Anish


___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] A couple of thoughts about moving forward.

[George Hunt]

On Sun, Aug 11, 2013 at 9:47 PM, David Farning dfarn...@activitycentral.com
 wrote:

 Hey all,

 I would like to offer some reflections after the last couple of weeks.

 I stepped aside because I felt I was hindering the project more than
 helping it. I spent years being frustrated by Langoff's hold on
 OLPC-XS. Then after less than 8 months I found myself controlling the
 funding for the 6 person DXS team, creating the roadmap  project
 specification, and doing much of the external communication. All of
 this while receiving dozens of emails and calls per week from
 deployments pressuring me to make XSCE and DXS different from what the
 core XSCE team was interested in doing.


My shortcomings may have caused a split between  XSCE and DXS.  When David
and I were discussing whether Ansible should be part of 0.4 XSCE, I felt a
fear of creating a situation I have created many times before, in my life
as a programmer. I tend to add more complexity than I have brain power to
sort out during the debugging phase.

So, now David has moved forward with DXS, with an aggressive schedule,
adding features based upon customers requirements. And when he wants to
incorporate DXS into the next revision, XSCE 0.5, the fear crops up again.
I need help dealing with my fear of complexity. Are there any volunteers?

In a sense it's the Red Hat, Fedora situation with a twist. The quick
turnaround, feature development test bed, is the commercial enterprise. The
volunteer, community based, effort is the slower moving, and more
conservative.

So now, our history, becomes our handicap. XSCE has not asked for much help
from the people and the accumulated wisdom available on server-devel. But
now I think we need that perspective.

I don't want to have a hold on XSCE. I'm feeling like I need to pass the
baton to someone, or a group of someones.  I've been working hard at a
volunteer job, and there just are no more hours in the day that I'm willing
to devote to the XSCE enterprise.

George


 It was not a recipe for community success :( So, I spent the last
 couple of weeks regrouping. If anyone has any suggestions for how they
 think I can help the community without becoming too smothering please
 let me know.

 I have been a little concerned about the relationship between the XSCE
 team and the DXS team. We put a pretty intense deadline of mid Oct for
 delivering commercially supported Dextrose Server. The goal of this
 division was to ensure the upstream XSCE team had the freedom to
 scratch their own itches while ensuring the downstream DXS team was
 focused on specific customer requirements. As a side effect it feels
 like there has become a gap between the teams.

 I would like to encourage Anna to step into the role of liaison
 between the two team. She can make sure that everyone is aware of what
 is happening.

 External communications hit a couple of rough patches over the past
 couple of weeks. While keeping the signal to noise ratio high, the use
 of a semi-private mailing list seemed to be hindering external
 awareness of what we were doing. Rather than ask the project to
 change, I decided to unsubscribe from this list and only remain
 subscribed to the server-devel list. The goal was to see how the
 projects was seen from the outside.

 My takeaway is that we should start to shift as many technical threads
 as possible to server-devel, there is a wealth of knowledge on that
 list. On planning and organization issues, the noise(passion) on
 server-devel might still might be a bit high for a young community
 like XSCE to handle without getting bogged down. I would suggest
 revisiting this decision one month prior to the release of 0.5.

 Good work everyone. Adolescents is a tough time for everyone
 especially community projects :)

 --
 David Farning
 Activity Central: http://www.activitycentral.com

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Some queries regarding PATHAGAR workflows

[George Hunt]

Hi Ajay,

As you start playing with pathagar on XSCE, could you look at what will be
involved to get it to work properly when the package is not located at the
root of the server, ie. when WSGIalias is set to /books/ or /pathatar/?

Seth looked at the access log,(copied below) and concluded that there would
be some programatic changes required.

Following is part of our conversation:


Seth wrote:
Django is definitely messing up.  books/book/add is the url you want,
right?  You are prefixing a url with /books/$pathagar.

If you can put pathagar at the root of your domain, it should just work.
 If hosting pathagar with a prefix is necessary there are two things we can
do:

(On the XSCE server, I think we need a prefix -gh)

1. Look up the right Django way to make this abstract.  I am pretty sure a
preferred way exists, but it will take me some time to seek the solution
out and implement it.

2. Modify all of the url routes in urls.py

  (r'^tags/(?Ptag.+)/$', 'pathagar.books.views.by_tag',

would become

  (r'^books/tags/(?Ptag.+)/$', 'pathagar.books.views.by_tag',

This is probably your best bet given your time constraints.

--Seth


On Sun, Aug 25, 2013 at 6:58 PM, George Hunt georgejh...@gmail.com wrote:


 After logging in as an administrative user to pathagar,  the following log
 snippet shows an effort to select a book to upload in the
 pathagar-error.log file:

 172.18.100.204 - - [22/Aug/2013:14:13:49 +] GET /favicon.ico
 HTTP/1.1 404 342 - Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_2)
 AppleWebKit/537.36 (KHTML, like Gecko) Chrome/28.0.1500.95 Safari/537.36
 172.18.100.204 - - [22/Aug/2013:09:13:54 -0500] GET /books/latest
 HTTP/1.1 301 - http://172.18.96.1/portal/; Mozilla/5.0 (Macintosh;
 Intel Mac OS X 10_8_2) AppleWebKit/537.36 (KHTML, like Gecko)
 Chrome/28.0.1500.95 Safari/537.36
 172.18.100.204 - - [22/Aug/2013:09:13:59 -0500] GET /books/latest/
 HTTP/1.1 200 3782 http://172.18.96.1/portal/; Mozilla/5.0 (Macintosh;
 Intel Mac OS X 10_8_2) AppleWebKit/537.36 (KHTML, like Gecko)
 Chrome/28.0.1500.95 Safari/537.36
 172.18.100.204 - - [22/Aug/2013:09:13:59 -0500] GET /favicon.ico
 HTTP/1.1 404 342 - Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_2)
 AppleWebKit/537.36 (KHTML, like Gecko) Chrome/28.0.1500.95 Safari/537.36
 172.18.100.204 - - [22/Aug/2013:09:14:08 -0500] GET /books/book/add HTTP/1.1
 200 6405 http://172.18.96.1/books/latest/; Mozilla/5.0 (Macintosh;
 Intel Mac OS X 10_8_2) AppleWebKit/537.36 (KHTML, like Gecko)
 Chrome/28.0.1500.95 Safari/537.36
 172.18.100.204 - - [22/Aug/2013:09:14:14 -0500] GET
 /media/js/admin/RelatedObjectLookups.js HTTP/1.1 404 369 
 http://172.18.96.1/books/book/add; Mozilla/5.0 (Macintosh; Intel Mac OS
 X 10_8_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/28.0.1500.95
 Safari/537.36
 172.18.100.204 - - [22/Aug/2013:09:14:14 -0500] GET /favicon.ico
 HTTP/1.1 404 342 - Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_2)
 AppleWebKit/537.36 (KHTML, like Gecko) Chrome/28.0.1500.95 Safari/537.36

 the book for upload was selected at this point, and the add button at
 the bottom right of the screen was activated

 172.18.100.204 - - [22/Aug/2013:14:14:26 +] POST /book/add HTTP/1.1
 404 339 http://172.18.96.1/books/book/add; Mozilla/5.0 (Macintosh;
 Intel Mac OS X 10_8_2) AppleWebKit/537.36 (KHTML, like Gecko)
 Chrome/28.0.1500.95 Safari/537.36
 172.18.100.204 - - [22/Aug/2013:14:14:26 +] GET /favicon.ico
 HTTP/1.1 404 342 - Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_2)
 AppleWebKit/537.36 (KHTML, like Gecko) Chrome/28.0.1500.95 Safari/537.36


 (color added to highlight the url which seems questionable-- second was
 different from the first)
 I believe that that the correct url to post the data would be
 /books/book/add

 This is because the httpd WSGIAlias in /etc/httpd/conf.d/pathagar.conf
 points to the pathagar.wsgi file, which I think also should be used as a
 prefix to the package.  Or--maybe I should be setting the prefix, and I'm
 not doing it

 As an additional observation, navigating to /books/ redirects to
 /latest/ rather than /books/latest, and returns a 404 Not found error.
 This may, or may not be related.


Ajay, are you testing on a physical XO, or are you using a virtual machine?

I'm glad to have your help resolving the remaining issues with pathagar on
XSCE.

George

On Fri, Sep 6, 2013 at 9:33 AM, Ajay Garg a...@activitycentral.com wrote:

 Thanks Aneesh for the pointers; I will keep them in mind when I start
 playing with PATHAGAR in XSCE :)
 Thanks again.


 On Fri, Sep 6, 2013 at 3:09 PM, Aneesh Dogra lionane...@gmail.com wrote:


 On Sep 6, 2013 1:06 PM, Ajay Garg a...@activitycentral.com wrote:
 
  Hi all.
 
  I have been playing with the public book-server
 http://108.171.173.65/latest/ that Sameer let me know some time back
 (thanks a lot Sameer, again !!)
 
  However, I have some queries regarding some workflows (and I found it
 best to query all you guys, so that the maximum information could be
 shared) :)  ::
 
 
 
  1

[Server-devel] rc.local stall on reboot

[George Hunt]

I'm in Haiti, with limited internet access.  Adam reports that the reboot
at Silar's orphanage is stalling, and when he hits escape, we sees a red
star traveling left and right, and an error message, which talks about a
stop script.


I am using rc.local to start up wvdial , which in turn stimulates the 3g
ZTE USB modem, and opens up a vpn connection.

What is the protocol so that it does not try to run the script also for
shutdown?

Any suggestions.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] A couple of thoughts about moving forward.

[George Hunt]

Hi Anish,

I look forward to playing with the XSCE installed via Ansible.

Will there be an install procedure, and cookbook, to try it out?

George



On Wed, Aug 28, 2013 at 11:37 PM, Anish Mangal an...@activitycentral.comwrote:

 Well, I was sort of hoping:

 
 * We could start to have discussions and work around some/all of the
 topics as a community. Everyone here has way more expertise than me in many
 (if not all) of the topics I listed. We can build a much better server if
 we all can use our expertise in the relevant part of the server. This
 transcends the pure software-development aspect of XSCE.


 
 * As the 0.4 version of the XSCE is nearing release, it's a good time to
 start thinking about additions/changes for 0.5. One of the consistent
 efforts (and demands) has been to make the server code more manageable, and
 by extension, modular and scalable.

 * 0.2.1 was a drop-in replacement of the XS-0.6/7
 * 0.3 involved a major reorganization to make the services more modular
 * 0.4 built upon that, by providing all the code in the same modular
 structure

 Within Activity Central, a team of developers (Santi, Miguel, Ajay, Anna)
 have been working on converting services available on the XSCE into
 ansible playbooks. The playbooks are written in a syntax which is *very
 easy to understand*, and the same playbook can be run on different
 platforms to produce the same effect. The playbooks can provide variables
 which may be integrated easily with other administration web-services (for
 example ajenti http://ajenti.org/).

 I hope to share the code for the playbooks very soon, so anyone can have a
 look at and try them. We have been able to get a fully functional server up
 just by playbooks and reusing/restructuring the available XSCE (xs-config)
 code.

 As someone leading the Dextrose Server initiative, I would push for the
 inclusion of these playbooks in XSCE-0.5. There is long term value in
 learning a bit of ansible and being able to work at a higher abstraction
 level.

 Best,
 Anish


 On Wed, Aug 28, 2013 at 6:48 PM, Tim Moody t...@timmoody.com wrote:

   Thanks for making this public.  What do you see as the next step?
 
 From: Anish Mangal
 Sent: Wednesday, August 28, 2013 6:16 PM
 To: xsce-de...@googlegroups.com ; server-devel ; Tim Moody
 Subject: Re: [XSCE] A couple of thoughts about moving forward.
 
 Hi Tim, et. al.,
 
 Since it was requested that I share my conversations with various
 deployments over the summer yielded in form of potential requirements for
 the school server, I created this wiki page:
 
 https://sugardextrose.org/projects/xsce/wiki/Primary_considerations
 
 
 There's obviously more data available, but what you see is a filtered
 version of guidelines I think we should keep in mind while developing a
 school server.
 
 Do the points in there (summarized below) make sense? I intentionally
 created this page on the sugardextrose.org wiki. If it has greater
 acceptance community-wide, I'd be happy to move it to the main XSCE wiki.
 
 * Statistics
 * Content
 * Internet traffic shaping
 * Administration
 * Networking
 * Classroom and School management
 * Total Cost of Ownership
 * Power
 * Sneakernet - LAN - Internet
 * i18n
 
 Best,
 Anish
 
 

 --
 Sig inserted by AutoHotkey ver. 1.1.11.01 (signature - first line)
 WLMail QuoteFix - http://www.dusko-lolic.from.hr/ (signature - second
 line)



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Western Digital hard disk data

[George Hunt]

James, et al,

Monitoring the transient voltage minimums, running on battery, not
charging, with 2 ethernet dongles as additional load I get the following in
my testing of WD external  my passport drive:

XO4 -- voltage droops to 4.51 -- spins up reliably
XO1.75 -- voltage droops to 4.56 -- spins up reliably
XO1.5 -- voltage droops to 3.97 -- does not spin up

On the XO1.5, when I used external powered hub, idle voltage was 5.2, which
drooped to 4.97 during spin up.

Then I used my newly purchased lab power supply to try to find the
threshold for the decision to spin up. When I adjusted it to 4.3, and my
meter registered a transient minimum of 3.96, it failed to spin up. When I
set the PS to 4.35, and the droop went down to 3.98, the drive spun up. I
judge that the threshold is pretty close to 4.0v.

George


On Tue, Aug 13, 2013 at 9:14 PM, James Cameron qu...@laptop.org wrote:

 What did the bus voltage do during these events?

 A HD would probably decide not to spin up if the bus voltage was out
 of spec.

 On Tue, Aug 13, 2013 at 02:00:31PM -0400, George Hunt wrote:
  Hi all,
 
  Following James' suggestion, I put my fluke 186 multimeter in current
 mode on
  the USB cable between the power source and the WD hard disk.  I was able
 to
  capture a 1A current spike, and a ~ 3 second current transient of about
 800mA
  during spin up.  The idling current for the hard disk, while spinning,
 was
  230mA.
 
  The XO4, (sku306) did not reliably spin up the HD. In particular this
 failure
  was almost 100% when the 430mA load of the ethernet dongle, and MR3020
 wifi
  access point was hanging on the other USB port.
 
  I put the hard drive and wifi accessories on a powered Plugable model
 USB2HUB
  AG7, which has an adapter rated for 3A, and experienced no failures,
 during
  many startups.
 
  I am concluding that an external hard disk requires a powered USB hub.
 
  Other interesting measurements:
 
• The MR3020 wifi unit had maximum current spikes of 530mA with average
  current of 175mA.
• The USB ethernet dongle had 230mA spikes with an average of 177mA
 and a
  minimum of 172mA
 
 
  George

  ___
  Server-devel mailing list
  Server-devel@lists.laptop.org
  http://lists.laptop.org/listinfo/server-devel


 --
 James Cameron
 http://quozl.linux.org.au/

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Re: Root fs on XO1

[George Hunt]

On Sun, Aug 11, 2013 at 6:49 PM, James Cameron qu...@laptop.org wrote:


 Rather, I guess that our dracut-modules-olpc package simply hasn't got
 scripted support for swinging the kernel over to a root filesystem on
 the SD card on XO-1.

 http://dev.laptop.org/git/users/dsd/dracut-modules-olpc/


 http://dev.laptop.org/git/users/dsd/dracut-modules-olpc/tree/30olpc-boot/olpc-boot-cmdline.sh

 The above script is where the decision is made, and XO-1 SD card is
 not in the case statement, because we didn't add this support.

 So yes, specifying a root filesystem using the command line may be a
 workaround.


Really helpful information.



  Then I started following James' suggestion, looking into
  olpc-os-builder.   There's an sd_image option which can be
  enabled.  I could not determine if enabling that would also enable
  booting from an external USB flash/hard disk, but it seems worth a
  try.

 The lack of support in dracut-modules-olpc would still need fixing, in
 my opinion.

 The sd_image option will make a .zd but XO-1 doesn't have fs-update.

  (I'm still toying with the idea of a WD Passport -- external hard
  disk as the supported option for the XO1 XSCE).

 I've avoided that brand because of extra features of the product, in
 case they interfere.  Does it work okay?


I was really fortunate, in my testing yesterday.  I happened to have a USB
ethernet dongle, and a 5v tplink 3020 wifi unit attached when I started
testing.  The WD disk did not spin up.  I removed the ethernet and wifi
loads, and it spun up correctly. The fortunate part was that a
vulnerability was exposed, and the opportunity to address it is at hand.

I'm going to spend some time today to get numbers for the interfering
loads. The XO4 under test is connected to a watts up power meter. This
morning, idle, the watts-up indicated 3.3w. I did an ls on the mounted
drive, and saw a transient on the meter to something between 7 and 8 watts.
 I'm sure I, and the meter, didn't see the whole event. But after a length
of time (probably spin up) the power was 4.6 (maybe the additional cpu and
spinning current). Letting the XO4, and the hard disk, go to idle, I
removed the hard disk USB connector.  After a short time, the power
decreased (after  a long time at idle) from 4.0w to 2.2w.

So the watts-up is probably not very accurate, but it yields this set of
observations:

   - The spin up creates a transient on USB power of almost an amp.
   - The spun down power,on idle, drawn by the WD hd is about 2 watts
   (4.0-2.2watts; =1.8w / 5V = .36A)
   - Failure to provide stiff enough power on USB 5v leads to HD
   initialization failures
   - Whether in the WD drive itself, or the OS, there will be spin down/up
   events during normal server operation
   - Adding back in the wifi adapter, and ethernet dongle raised the idle
   power to 5.6 Watts (from 4.0w).
   - The power to keep the disk spinning is apparently about .6w.
   - When I disconnected the tplink mr3020 there was a .6w change in power
   = .6w / 5V = .12A = 120mA (idle--no RF activity)


So I ordered 2 12v USB hubs from amazon to do some additional testing:
http://www.amazon.com/Anker%C2%AE-Uspeed-4-Port-Adapter-Chipset/dp/B005QWY3PU
http://www.amazon.com/StarTech-com-Mountable-Rugged-Industrial-ST4200USBM/dp/B001H7AFO6/

I had also ordered, earlier, the following items, which I think might be
needed in a deployment where the school server needs to be available during
the school day, even when the power happens to be off:
http://www.amazon.com/gp/product/B003QRQ5DK/
http://www.amazon.com/gp/product/B00A71CMDU
http://www.amazon.com/gp/product/B003UANXCE/

I'm beginning to solidify in my mind a system which includes some or all of
the above, which I can begin to test. I've added to the list a Sears deep
cycle battery, which I bought locally for about $130.

I had concluded, earlier, that the loop resistance of a 300 ft max cat5
ethernet hub-to-hub link is about 8 ohms. With the 120mA observation above,
for the power consumed by the tplink mr3020, we'd expect a .96V supply
droop running mr3020's at max distance -- not likely to work.  That's why I
think running power-over-the-internet at 12v with the 12v dc to 5
converter, seems attractive.  But it will take more experimentation with
mr3020's to see how much supply droop they can handle -- and where the
cable length begins to require dc conversion.

(Extra features of consumer backup drives _might_ involve changes to
 the device in terms of the USB descriptors, such as those required for
 a push button to start backup software, or CD-ROM images of the
 software, so I've tended to select drives that either don't mention
 the software, or are marketed as expansion rather than backup.)

  I'm not yet clear whether it would be any simpler to insist that
  XO1's be unl

ocked before installing XSCE. It's easy to modify the
  kernel command line via olpc.fth, (which I verified to be ignored
  on a locked XO) but maybe getting the right 

Re: [Server-devel] [XSCE] Re: Root fs on XO1

[George Hunt]

Thanks everybody for your responses.

I learned a lot yesterday by trying things you all suggested.  James' idea
of using a XO1.5, and devalias. . .,fs-update . .  to populate an SD card
was quick, and booted on a 1.5.  I tried copying the /boot directory of a
XO1 to this SD, and it appeared that the signed kernel was recognized,
loaded, but dracut issued a warning that root argument was empty.

So then I read up on dracut. My guess is that the drivers to read the SD
card were missing from the initramfs.

Then I started following James' suggestion, looking into olpc-os-builder.
 There's an sd_image option which can be enabled.  I could not determine if
enabling that would also enable booting from an external USB flash/hard
disk, but it seems worth a try. (I'm still toying with the idea of a WD
Passport -- external hard disk as the supported option for the XO1 XSCE).

I'm not yet clear whether it would be any simpler to insist that XO1's be
unlocked before installing XSCE. It's easy to modify the kernel command
line via olpc.fth, (which I verified to be ignored on a locked XO) but
maybe getting the right drivers in the initramfs is the hard part.  I still
have lots more to learn.

George


On Fri, Aug 9, 2013 at 9:30 PM, Mikus Grinbergs mi...@mikusj.com wrote:

 On 08/09/2013 04:29 PM, Jerry Vonau wrote:

 Are you boot from the SD card, or booting from internal storage and using
 the SD card as more storage?


 I have not had reason to boot from SD card - so I've never done that. [I
 have on occasion booted from USB stick, in order to repair something in
 jffs2 on the XO-1.]

 After I install a new build on an XO, I run (as root) some scripts which
 link in the resources on the 'permanent' SD card to the operating system
 on the XO (and then I reboot).  In my case, ALL 'permanent' SD cards (for a
 given version of Fedora, and a given architecture) have the same content.

 I had difficulty in the past in making the resources on the SD card
 available in time for Sugar to start up.  Accordingly, I'm using a script
 in /etc/rc.d/init.d to ensure the SD card is mounted.

 mikus


___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Root fs on XO1

[George Hunt]

Hi Jerry, et al,

The motherboard flash on an XO1 is 1GB. The kernel, rootfs, provided by
OLPC Boston for the XO1 occupies 745MB. The server software (XSCE) we have
been adding on top of the OS occupies about 1.3GB.  So obviously an
additional SD card is required.

My first approach has been to start with a vfat formatted SD card, change
its format to ext4, without changing its partition layout (I think this
avoids creating erase block/OS write mismatch, which slows and wears SD
cards prematurely). Then I have been copying large chunks from the JFFS to
the SD card, and sym-linking to it, so that additional yum install
operations are diverted to the SD card.

Recently, Jerry has suggested that I just ignore the motherboard jffs
flash, and run entirely off the SD card.  I've been studying how to
accomplish this.  It looks to me like the OS.img file available for the
XO1 is in jffs2/mtd format which would be appropriate for dding directly
to the motherboard flash, but probably not correct for dding to an SD
card.

So here is my question: Does it make sense to let the openfirmware bios
write the OS image to the jffs on the motherboard, and then for me to rsync
that OS, as data,  to an ext4 formatted (not repartitioned) SD card? If I
do this will the presence of a signed kernel on the SD card cause the boot
loader to choose the SD card, ignore JFFS, even if the XO is still secured?

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] XSCE Update day 3

[George Hunt]

Today, we are focusing on integration of Internet In a Box, the rpm's, and
the 600GB  of external hard drive data and hardened up objectives for the
rest of the week.

We discussed French/Creole/Spanish translation of IIAB GUI?  Braddock will
try to send the English phrases (several hundred) to Holt next week
sometime, so http://translate.sugarlabs.org might even support wider
translations.

The data set Braddock sent to George last week is not a final release.
Braddock will forward updates, or a new drive, by Aug 15 hard drive in NYC
/ Aug 25 hard deadline (flight!) for Haiti.  Aug 5 XSCE 0.4 target: is
http://schoolserver.org/0.4/Road_Map real?

Today's discussion developed a more detailed punch list of features,
responsibilities as we move forward to the feature freeze Aug 5: (following
is cut and paste of meeting minutes)


   1.

   Specific scope, testing, and documentation for 0.4.
   1.

  Offline USB install -- Jerry: GO/NOGO
  1.

 Next Steps: Working on finals script.
 2.

 Documentation:. Not yet
 3.

 Testing: ?
 2.

  Install image -- Jerry: GO/NOGO -- promising -- it  boots?
  1.

 Document: Not yet
 2.

 Testing: Not yet
 3.

  External USB hard drives (currently have power issues) -- Not
  Assigned Yet: GO/NOGO verify on 13.2. This might be hard drive
  specific. Must recommend type of hard drive/frame or external drive.

  Testing may actually occur during testing of IIAB (item d.)
  4.

  Internet in a Box -- George: GO/NOGO
  1.

 Documentation: Not Yet
 2.

 Testing: Not yet
 5.

  Pathagar -- Lion: GO/NOGO
  1.

 Documentation: Not Yet
 2.

 Testing: Not yet
 6.

  Ansible --Santi: GO/NOGO don’t comment out from set up until jerry is
  done with offline install?
  7.

  WebDAV to submit work from the Journal -- Ajay: GO/NOGO design docs.
  use case?
  8.

  Monit -- Santi: GO/NOGO
  1.

 Documentation: Not Yet GO
 2.

 Testing: Not yet
 9.

  Dansguardian modularity (currently it's entwined with Squid ports) --
  -- need to be able to disable dansguardian without disabling sqid. Santi:
  GO/NOGO
  10.

  Performance Statistics Collection -- George/Anish:  GO/NOGO
  11.

  Activity Update (separate from the Activity Server) -- ???: GO/NOGO
  1.
  12.

  Content updater -- Jerry: GO/NOGO
  13.

  Printing -- George: GO/NOGO
  1.

 Include rpm but don’t enable.
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [support-gang] XSCE Sprint

[George Hunt]

Thank you Tony for you perspective. So can we begin to look at
implementation details?

Perhaps discussions about how to implement new features belong on
server-devel rather than support gang.  My apologies if so. But if XSCE
begins to offer features that are really used and needed in the classroom,
then the meaning of support will start changing.

George

On Tue, Jul 9, 2013 at 10:43 AM, Tony Anderson t...@olenepal.org wrote:

 Hi, All

 I am sure this is a broken record by now, but I keep hoping that there
 will be focus on content. I am not immediately concerned about power
 issues. The move from 386ISA to Arm has good potential, but the stability
 and functionality of XS-0.7 makes it still the software of choice. By
 mid-2014, the new generation of Atom may provide even lower power
 consumption.

 The IIAB (internet In a Box) now mandates a TB+ hard drive. This is viable
 for Atom based systems. It is also viable on the Trim-Slice H. My concern
 with the Trim-Slice is the 1GB memory. A server for a school of 100+
 laptops probably needs more. 2GB is good, 4GB is better.


There have been recent conversations with the Trimslice manufacturer, to
get a version which has 2-4GB memory, quad core processor, and 2 ethernet
connections.


 My urgent concerns are:

 1 - An effective way to organize the digital library so that kids are
 attracted to find items they would like to download.


We hope that we can evolve, and incorporate Pathagar for this in the short
term.  I've asked if there are other  open source alternatives, and not
gotten any viable suggestions. Seth insists that Pathagar is only going to
work for books. I'd like a multimedia warehouse.


 2 - Provide for a shared printer attached to the server which serv- es all
 of the XOs but gives the teachers and administrators control over the use
 of expean andible resources.


Just looking at low hanging fruit --  What do you think of a PHP file on
the XSCE web server to initiate a file upload, and using Browse Activity to
extract Journal entries? We could have these uploaded files dropped into a
directory where the teacher could trigger a print job.  There may be  a
client/server interface in CUPS which lets the teacher administer, and
trigger print jobs, from her own laptop.


 3 - an implementation of Puppet or similar technology to allow update of
 the XOs - supporting mix of XO-1 to XO-4, providing for reflash as well as
 updates (something like Nandblaster). This should work equally well for
 locked and unlocked XOs.


I'm not sure how to achieve these objectives. We have been exploring
another package very similar to puppet, ansible, might be able to achieve
the same outcomes, and be simpler to administer.  Reflashing seems very
different from in-place upgrading.

Both puppet, and ansible, require root access permissions, and a
functioning operating system.  If the target machine has broken software, I
don't see an easy alternative than to reflash with a signed image, just to
verify the hardware.



 4 - some way for the school server to support email, proabably a
 light-weight client Sugar activity and postfix or similar on the server.
 There needs to be a mechanism to offload pending email for the outside
 world to a usb drive which can be sent at a cyber cafe as well as a way to
 receive email at a cyber cafe for insertion into the school server queue.
 It would be nice if the kids could have a school-based email address but
 this means setting up an internet domain somewhere. Gmail would work, but I
 am not sure how that can be done on the LAN when not connected to the
 internet.


This need is not in my area of expertise. Sendmail is often installed by
default, but is complex to configure. Does anyone have recommendations on
how best to achieve this objective?



 5 - upgrade to Moodle 2 along with some strategy on how it could be used
 in the local school environment. My approach is to use it for a 'teacher's
 lounge', a place for teachers to find lesson plans, training materials, and
 a method to communicate with each other. In reality, teachers do not have
 time to take advantage of this. Sadly, even though Moodle is open source,
 there is very little open source courseware available for Moodle (one good
 digital photography course). Naturally, Moodle has the additional
 disadvantage that courseware can not be downlaoded to the XO so that
 students can work at home. Moodle assumes the student is online.


We spent a little time looking at the resources required, server
horsepower, etc. and decided that moving to a more recent moodle was not
 an easy choice to make


 6 - support for GSM modems and wifi dongles on the school server. This is
 minor, but network configuration for this has to be done at the command
 line (no gui network manager). More important is a model for 'ET call home'
 for the school server. I believe in our deployments, the internet model is
 going to be more 'batch' jobs than online surfing. We all have experienced

Re: [Server-devel] [support-gang] Value of remote access to School Servers.

[George Hunt]

The central openvpn server would be configured to pass out local
(unroutable in the wider internet) addresses in the 10.0.0.0/8 subnet to
each client.

There would be one public/private key pair distributed with the XSCE
software distribution, for testing. The server would be configured to
accept multiple conections from the same key pair. Effectively this would
create a party line', where everyone who had access to the key pair, would
have access to the party line.  Then they would be able to ping all the
other XSCE servers, on the local 10.0.0.0/8 virtual private network (which
is worldwide) -- assuming that the firewalls were set to enable ping
responses. And they could log into any servers on that party line, for
which they had ssh authentication credentials.

Then, most likely with passwords turned off, deployments could use
public/private key pairs they generate themselves to access their own
servers.

For an additional level of security, deployments could contact
activitycentral to get their own public/private key pairs, one for each
machine, and a config file which connects to different ports, openvpn
instances, virtual box instances, or whole physical machines.

At the extreme, a deployment could have it's own virtual private network,
protected by key pairs known only to itself, on it's own machine, running
under lock and key, in its own back room, and then ssh (password or key
pair) connection to each of its machines.

George
George


On Wed, Jul 3, 2013 at 4:36 AM, Anish Mangal an...@activitycentral.comwrote:



 On Wed, Jul 3, 2013 at 1:54 PM, James Cameron qu...@laptop.org wrote:

 On Wed, Jul 03, 2013 at 12:45:35PM +0530, Anish Mangal wrote:
  James wrote:
   Would the person accessing their XSCE remotely then establish
   another tunnel to your OpenVPN server, or would your server do
   inbound connection forwarding?
 
  Hmm. I'm not so clear on that. I can give the example of a setup in
  Bhagmalpur (a pilot we recently did).
 
  1. There is an openVPN server hosted by Sameer.
  2. The XSCE when connected to the internet dials into this open vpn
 server.

 Thanks, I understand the first two steps, and they sound good.

  3. I can login to the XSCE through the openVPN connection through
 ssh and administer remotely.

 How is this last step achieved?  There's much flexibility, so I'm
 curious.  I imagine one of three methods:

 a.  does the user first SSH into an account on the OpenVPN server and
 then SSH again to the XSCE, or;

 b.  does the user SSH to a particular port on the OpenVPN server that
 is automatically forwarded to the XSCE, or;

 c.  does the XSCE have a routable IP address, courtesy of the OpenVPN
 server, to which SSH is directed?


 I'm not sure... let me explain (perhaps Sameer or Santi can chime in)...

 I have a set of openVPN keys on may laptop through which I connect to the
 openVPN server automatically (and a network called tun0 is created)

 I know the IP address of the XSCE in Bpur

 So, from my laptop, I just do ssh root@ip address of XSCE on the openVPN
 network

 Does it make things any clearer?


 --
 James Cameron
 http://quozl.linux.org.au/



 ___
 support-gang mailing list
 support-g...@lists.laptop.org
 http://lists.laptop.org/listinfo/support-gang


___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [XSCE] Fwd: XSCE links to filesfrom Moodle

[George Hunt]

David,

It sounds like moodle (a php application running under the httpd server,
which is itself running as the user apache) does not have read permission
on the files you would like to serve.  I checked my instance of XSCE. My
/library folder has root:root user and group with read and execute
permissions for (owner,group,world) -- (rwxr-xr-x). All the folders in the
path from the root directory to the content you want to serve must have the
x, and the files you want to serve must have permissions rw-r--r--
(readable by world- or you could change group to apache, and make them
readable by group).
You can insure that the files themselves are readable by world with the
command chmod o+r *

I'm guessing that you did not put your content at the root level of
/library, but that you created a /library/content directory. If so, you
might need to make the /library/content directory executable, as well as
readable by world -- rwxr-xr-x which is achieved by  chmod 755
/library/content

Pardon, if I made the incorrect assumptions, or said things you have
already tried,

George


On Wed, Jun 26, 2013 at 1:19 AM, Adam Holt h...@laptop.org wrote:

 Anna, any clues if you're awake??


 From: David Leeming da...@leeming-consulting.com
 Date: Wed, Jun 26, 2013 at 12:36 AM
 Subject: [Server-devel] XSCE links to filesfrom Moodle
 To: server-devel server-devel@lists.laptop.org


  Hi,

 ** **

 With the XS 0.7 we add links to Moodle pages that map to locations in the
 storage folder /library using aliases and html.conf file

 ** **

 I have gone through the exact process as we used to use with the XSCE and
 all good in the set up, but when I try to access those resources with a
 connected laptop or authenticated moodle account t it tells me that I don’t
 have permission to view the folder.

 ** **

 A missing step I think...

 ** **

 *David Leeming*

 Solomon Islands 

 24419 (h)

 www.rurallink.com.sb

 ** **

 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] External hard disk testing for use with IIAB (internet in a Box) on XO based XSCE servers

[George Hunt]

I have a number of 2.5 SATA drives, that I've used for I don't know how
long.

I put one in a startech.com external hard drive encosure model sat2510u2E.

I wrote a little cron stimulated script to write, and check for file
existence, and almost immediately got failures. I was using a USB hub
(plugable model:USB1-HUB-AG7) with a 5v 3A wall wart powering the unit
under test.

Googling the ehci_hcd failure messages, I find that there are a number of
longstanding issues in running these external enclosures under linux
kernels, even back in 2007 and earlier.

It's not clear to me yet whether these problems are disk or interface
related.

So I got paranoid, and connected a western digital drive (3TB self
contained 3.5in SATA USB interface with wall wart PS) that has been working
for me for almost 2 years. No failures in 24 hours.

Then I started learning about S.M.A.R.T., which is a disk self monitoring,
reporting, that most modern disk incorporate.  Running smartctl, which is
part of our OS, on my old disk, and the one which exhibited failures,
indicates that the drive itself does not know that it is failing.

I'll probably modify the included script to use smart, in the next
iteration.  But I thought I'd fish for any expertise within the olpc
community.

Are any external enclosures known to work? I have't been able to find
descriptions of the chipsets that do the USB interface.

George

Maybe this is too simplistic:


#!/bin/bash
# script to check for external hard disk presence and writeability
MOUNTPT=/mnt/usb0
# read a file that is always there
dtm=`date`
if [ ! -f $MOUNTPT/token ]; then
echo $dtm   failed to read $MOUNTPT/token. Writing it /root/disk.log
echo this is content written at $dtm  $MOUNTPT/token
else
echo $dtm -- success so far  /root/disk.log
fi
# check for the absence of a file
if [ -f $MOUNTPT/writeit ]; then
echo $dtm -- last attempt to remove $MOUNTPT/writeit failed 
/root/disk.log
else
echo $dtm -- content of file writeit  $MOUNTPT/writeit
fi
sync
if [ -f $MOUNTPT/writeit ]; then
rm $MOUNTPT/writeit
else
echo $dtm -- failed to write $MOUNTPT/writeit  /root/disk.log
fi
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Best XSCE strategy for us in Pacific with XO-1.5 2GB Testing XSCE 3 on XO 1.5 2GB os855

[George Hunt]

I followed the suggestions at
http://wiki.laptop.org/go/Release_notes/10.1.0#External_SD --
substituting 31035o1.zd
for the earlier OS64.

devalias fsdisk /sd/disk@1:0
fs-update u:\os64.zd

after the reboot, the df command indicated rootfs had a size of 60GB. I
was able to write and read from the command line as olpc and as root. So I
believe that indicates that 64GB drives do work on the XO-1.5.

George


On Mon, Jun 17, 2013 at 10:55 PM, James Cameron qu...@laptop.org wrote:

 On Mon, Jun 17, 2013 at 09:20:20PM -0400, George Hunt wrote:
  I ordered a 64GB SD, and have it in hand. It is formatted with exfat
  disk format, apparently a Microsoft proprietary standard.  There is
  a driver for exfat, but it is not shipped with the XO, so there was
  some truth to the hearsay that 64GB SD cards do not work on the
  XO. BUT . . .
 
  I was able to read exfat formatted out of the box transcend 64GB
  drive on a XO1.5, with a downloaded driver, so I expect that
  electrical compatibility is already established.

 It is the protocol compatibility that causes us problems.  There's
 this tiny computer inside the card which has to interoperate with our
 computer.  And from what you describe above, you haven't yet tested
 the protocol compatibility with Open Firmware.  You have to boot from
 it to test that.

  I think you will want it to be formatted with ext4, because the OS
  doesn't run on a FAT formated drive. I should caution that I have
  read that SD cards have an erase block size that makes it important
  to preserve the partition location and sizing, as the disk format is
  being erased and reformatted.

 If you use the drive for the operating system, there's no need to
 partition or format, because the fs-update of 13.1.0 .zd file will
 create all that is required.  Just make sure you use the devalias
 fsdisk ext:0 trick to persuade fs-update to use the external slot.

 During boot, the operating system may expand the filesystem to cover
 the rest of the device.

  I do not expect that the prep-storage.sh script which is part of
  XSCE-0.3, will correctly format a 64GB SD card. I do not know
  whether the XSCE install process will properly function when it is
  loaded into a 64GB SD. (I expect probably so, since I don't think
  the OS would know whether it was running on the mother board or on
  an external SD card).

 It must know, and does know.

  I'll experiment with my new SD card, and try to document the results
  with a script that you can use to format big SD cards.

 I suggest gparted for partitioning and formatting.

 --
 James Cameron
 http://quozl.linux.org.au/

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Best XSCE strategy for us in Pacific with XO-1.5 2GB Testing XSCE 3 on XO 1.5 2GB os855

[George Hunt]

This is probably a question for James C. I'm concerned that I may have
invalidated the verification, or at least created misleading information.

My confession: I couldn't pass up the chance to evolve our
prep-storage.sh script that is part of XSCE before I lost the exfat
formatting. So I tweaked on our script until it recognized and dealt
correctly with exfat format.  So when I did the fs-update os.zd, the
64GB SD card was already formatted ext4.

Do you think that a separate formatting step is necessary, or irrelevant,
to the success of fs-update?

George


On Tue, Jun 18, 2013 at 12:27 PM, George Hunt georgejh...@gmail.com wrote:

 I followed the suggestions at
 http://wiki.laptop.org/go/Release_notes/10.1.0#External_SD --
 substituting 31035o1.zd for the earlier OS64.

 devalias fsdisk /sd/disk@1:0
 fs-update u:\os64.zd

 after the reboot, the df command indicated rootfs had a size of 60GB. I
 was able to write and read from the command line as olpc and as root. So I
 believe that indicates that 64GB drives do work on the XO-1.5.

 George


 On Mon, Jun 17, 2013 at 10:55 PM, James Cameron qu...@laptop.org wrote:

 On Mon, Jun 17, 2013 at 09:20:20PM -0400, George Hunt wrote:
  I ordered a 64GB SD, and have it in hand. It is formatted with exfat
  disk format, apparently a Microsoft proprietary standard.  There is
  a driver for exfat, but it is not shipped with the XO, so there was
  some truth to the hearsay that 64GB SD cards do not work on the
  XO. BUT . . .
 
  I was able to read exfat formatted out of the box transcend 64GB
  drive on a XO1.5, with a downloaded driver, so I expect that
  electrical compatibility is already established.

 It is the protocol compatibility that causes us problems.  There's
 this tiny computer inside the card which has to interoperate with our
 computer.  And from what you describe above, you haven't yet tested
 the protocol compatibility with Open Firmware.  You have to boot from
 it to test that.

  I think you will want it to be formatted with ext4, because the OS
  doesn't run on a FAT formated drive. I should caution that I have
  read that SD cards have an erase block size that makes it important
  to preserve the partition location and sizing, as the disk format is
  being erased and reformatted.

 If you use the drive for the operating system, there's no need to
 partition or format, because the fs-update of 13.1.0 .zd file will
 create all that is required.  Just make sure you use the devalias
 fsdisk ext:0 trick to persuade fs-update to use the external slot.

 During boot, the operating system may expand the filesystem to cover
 the rest of the device.

  I do not expect that the prep-storage.sh script which is part of
  XSCE-0.3, will correctly format a 64GB SD card. I do not know
  whether the XSCE install process will properly function when it is
  loaded into a 64GB SD. (I expect probably so, since I don't think
  the OS would know whether it was running on the mother board or on
  an external SD card).

 It must know, and does know.

  I'll experiment with my new SD card, and try to document the results
  with a script that you can use to format big SD cards.

 I suggest gparted for partitioning and formatting.

 --
 James Cameron
 http://quozl.linux.org.au/



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] ANNOUNCEMENT XSCE 0.3 Final Release

[George Hunt]

I just bought a 64GB transcend SD. I'll share the results.


On Wed, Jun 12, 2013 at 8:58 AM, James Cameron qu...@laptop.org wrote:

 On Wed, Jun 12, 2013 at 07:50:07AM -0500, Jerry Vonau wrote:
  On Wed, 2013-06-12 at 21:16 +1000, James Cameron wrote:
   On Wed, Jun 12, 2013 at 05:26:37AM -0400, George Hunt wrote:
It's my understanding that 32GB cards work, but that 64GB cards
do not (hearsay, and not direct personal experience).
  
 
  Think you ran across this issue before:
 
  http://lists.laptop.org/pipermail/devel/2012-October/036017.html
 
 
   I'm interested in working with anyone who has one, in order to
   diagnose and fix this.  Let me know.
  
 
  Think this has been corrected:
 
  http://dev.laptop.org/ticket/12216

 Yep, would need to work with someone with one of these cards to make
 sure.  I'd check versions first.

 --
 James Cameron
 http://quozl.linux.org.au/

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] 12 Volt power system for School Servers

[George Hunt]

I'd be interested in ddwrt on tp-link 2030's, primarily because of the
Nepali data, that they can handle 25 clients, and because they run on 5v.

Add to that a DC-DC buck converter to 5v at the remote end of poe. I'm
testing
http://www.amazon.com/gp/product/B00A71CMDU/ref=oh_details_o07_s00_i00?ie=UTF8psc=1which
in the spec has 8-23v input capability, and 3A output ($8)

Let's analyze the worst case for an ethernet cable at maximum distance. (24
awg wire at 70C has resistance of 29 ohms/1000 feet, poe has 2 wires per
leg, but we have drop in each leg. So my guess is worst case loop
resistance of 300 feet * 29 ohms/1000 feet * 2 legs / 2 wires/leg = 8.7
ohms loop resistance.

If the minimum input voltage for the DC-DC is 8, then we only have 3 volts
of drop to play with. The current input spec for the tp-link is 1A. So if
300 at 1A would dorp 8.7 volts, and we can only afford 3 v drop, we can
only extend the AP 100 feet from the server.  I suspect that the current
draw is much less than the spec,  worst case battery discharge,ambient
temperature, are not likely, so more than 100 feet is likely.

This solution, would probably work for 3 or 4 adjacent classrooms, and
75-150 students -- but needs to be reality tested.

George


On Thu, May 30, 2013 at 12:47 AM, server-devel-requ...@lists.laptop.orgwrote:

 Send Server-devel mailing list submissions to
 server-devel@lists.laptop.org

 To subscribe or unsubscribe via the World Wide Web, visit
 http://lists.laptop.org/listinfo/server-devel
 or, via email, send a message with subject or body 'help' to
 server-devel-requ...@lists.laptop.org

 You can reach the person managing the list at
 server-devel-ow...@lists.laptop.org

 When replying, please edit your Subject line so it is more specific
 than Re: Contents of Server-devel digest...


 Today's Topics:

1. Re: 12 Volt power system for School Servers. (James Cameron)
2. Re: 12 Volt power system for School Servers. (Sameer Verma)
3. Re: 12 Volt power system for School Servers. (David Farning)
4. Re: 12 Volt power system for School Servers. (Anna)
5. Re: 12 Volt power system for School Servers. (Adam Holt)
6. Re: 12 Volt power system for School Servers. (Tony Anderson)


 --

 Message: 1
 Date: Thu, 30 May 2013 08:26:32 +1000
 From: James Cameron qu...@laptop.org
 To: David Farning dfarn...@activitycentral.com
 Cc: server-devel server-devel@lists.laptop.org,   Tony Anderson
 tony_ander...@usa.net
 Subject: Re: [Server-devel] 12 Volt power system for School Servers.
 Message-ID: 20130529222631.gb20...@us.netrek.org
 Content-Type: text/plain; charset=iso-8859-1

 On Wed, May 29, 2013 at 04:51:47PM -0500, David Farning wrote:
  Any thoughts on powering the AP??

 They are all different.

 First, select an AP that says it has a 12V input.

 I would record the voltage without load emitted by the power supply
 supplied by the manufacturer.  This is a known maximum voltage that
 the AP can take.

 Then I would record the voltage on load.  This is a known minimum
 voltage.

 Then I would record the input voltage specifications of the access
 point using the manufacturer's documentation.  This is usually a
 typical voltage, but sometimes they include a maximum or minimum.  If
 there was no documentation, I'd ask them, though perhaps with no great
 expectation of success.

 Taking the minimum and maximum of those three values, I'd compare them
 to the lead-acid battery charge cycle operating range of 10V to 15V.

 If there remains any doubt (e.g. they haven't specified a maximum), I
 would attach the AP to a variable power supply and gradually ramp it
 up to 15V, checking for sudden drop of current (a blown fuse), or
 excessive operating heat.

 Once I'm happy, I would cut the cable as far from the device as
 possible, and reterminate it there.  Before cutting, I would discharge
 the power supply capacitors ... saves wear and tear on wire cutters.

 If the AP was USB powered Wifi dongle, the problem goes away.

 --
 James Cameron
 http://quozl.linux.org.au/


 --

 Message: 2
 Date: Wed, 29 May 2013 15:48:48 -0700
 From: Sameer Verma sve...@sfsu.edu
 To: James Cameron qu...@laptop.org,   David Farning
 dfarn...@activitycentral.com, Jerry Vonau 
 jvo...@shaw.ca,
 server-devel server-devel@lists.laptop.org,   Tony Anderson
 tony_ander...@usa.net
 Subject: Re: [Server-devel] 12 Volt power system for School Servers.
 Message-ID:
 CAFoGK8G3ZRQdj+1qyCiv8=
 putnsch5+7b4levahwr-zsxbb...@mail.gmail.com
 Content-Type: text/plain; charset=ISO-8859-1

 On Wed, May 29, 2013 at 3:26 PM, James Cameron qu...@laptop.org wrote:
  On Wed, May 29, 2013 at 04:51:47PM -0500, David Farning wrote:
  Any thoughts on powering the AP?
 
  They are all different.
 

 Yep. The one we use in Bhagmalpur (India) and in Jamaica is the
 Ubiquiti 

Re: [Server-devel] ds-backup

[George Hunt]

Your analysis conforms with my experience. When I tried to carry mod_python
over into FC18, it conflicted with apache. I didn't spend much time
thinking about carrying apache along.

I'm sort of looking forward to debugging my new version of ds-backup
server, which uses mod_wsgi. It seems like a relatively localized change.

George


On Tue, Apr 16, 2013 at 6:24 PM, Daniel Drake d...@laptop.org wrote:

 On Fri, Apr 5, 2013 at 3:16 PM, George Hunt georgejh...@gmail.com wrote:
  Hi Daniel,
 
  I noticed that you were working on ds-backup recently.  I was trying to
 get
  XSCE running on fc18, and encountered the issue of fedora dropping
  mod-python (see https://fedorahosted.org/rel-eng/ticket/5165).
 
  Is it ok to just carry the fc17 version along in our repo, and try to
 see if
  it just works, or should we set a goal of rewriting to mod_wsgi?

 Carrying along an old mod_python might be tricky, I think F18 has a
 new apache version with some incompatibilities.

 I would port it to wsgi.

 Daniel

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Journal backup on school server

[George Hunt]

Hi Tony,

I'm playing with ds-backup now, because one of the dependencies for
ds-backup is mod_python which was dropped from fedora 18, in favor of
mod_wsgi. It seems like a great time to rethink the issue.

I cannot find the working code you mention in this email in my google
stack.  Can you send it to me again?

I'm also talking with Tim Moody, about using puppet client as a way of
pushing and modifying the configuration of XO's in the classroom.  Do you
have any experience, or ideas, about such a proposal? I'm not sure whether
ds-backup might be used to introduce puppet client back into XOs.

George

On Sun, Dec 9, 2012 at 12:25 PM, Tony Anderson tony_ander...@usa.netwrote:

 Hi,

 I will send you the scripts as soon as I have a chance to try them on
 12.0.1 and XS-0.7. Last year they were running on xs-0.6 and build 852 at
 Saint Jacob. There shouldn't be too much difference, although so far on
 most of the code I have been wrong by about 3 days!

 The identity problem really isn't that hard. A lot of the mailing list
 discussion seems related to running Sugar on something besides an XO.

 The user is identified on the server by the serial-number. For the backup,
 this is not a problem. The problem comes from OLPC where an XO is shared.
 Then there needs to be something like a login to identify the actual user.
 There also needs to be separate Journals on the XO and separate backups on
 the school server.

 I should be able to give you working code within a week (I leave Kigali on
 12/23 so it must be before then).

 Yours,

 Tony


 On 12/09/2012 04:11 PM, George Hunt wrote:

 Wow Tony,

 Your solution definitely needs to be part of the mix as we go forward.
 Can you send me copies of the scripts, or the changes you made to
 accomplish these objectives?  If we are going to look at the
 serialnumber-user-identity issue, we might be making changes to the same
 packages.

 Have you worked through the changes needed to add user identity to
 journal backups? As Paul Fox was suggesting, I think there will be lots
 of ripple effects, if and when we start adding additional users, in
 addition to the user olpc.  Or maybe, someone will come up with a
 simplifying assumption, or approach.

 George



 On Sun, Dec 9, 2012 at 4:49 AM, Tony Anderson tony_ander...@usa.net
 mailto:tony_ander...@usa.net** wrote:

 Hi, Sameer

 I got an email from Nick Doiron re his visit to the Marshall
 Islands. He mentioned that you gave him a script to create a csv
 from the Journal backup.

 As I have mentioned several times on the list, I believe Martin's
 backup scheme while elegantly implemented is not adequate. His model
 is the traditional backup/restore.

 The problem, as always, is storage space. When an XO-1 is out of
 space, the user gets a message 'Journal is full'. The only practical
 solution often is to reflash the XO losing all of the Journal
 objects. If a user deletes an object from the Journal on the XO, the
 rsynch also removes it from the backup.

 I modified the scripts ds-backup.sh and ds-backup.py on the XO to
 use a different paradigm. The registration process creates two
 scripts in the /library/users/serial-number folder on the school
 server: journal and log. The backup script uploads the object to
 journal if there is an associated data file; otherwise, it is
 uploaded to the log folder and deleted from the local store. The
 remaining objects are marked as favorites in the Journal Activity to
 show that the data file is available on the XO. If the user clears
 the star, the data file is deleted from the datastore but remains on
 the school server. If the user fills a clear star, the backup script
 downloads the data file from the school server.

 At Saint Jacob, all of the laptops are 4gb XO-1.5 so I have not had
 to implement storage management. The plan would be to set limits on
 the size of the datastore and the /home/olpc/Activities folders. If
 a datastore outgrows the limit, data files can be deleted LRU.
 Similarly, activities can be deleted LRU. This way the user will not
 have to be involved. The user can always request a needed file to be
 restored to the local datastore and can request an activity be
 downloaded and installed from the school server repository. I expect
 this will be needed in Lesotho for XO-1.

 This method also supports restoring the users Journal after a reflash.
 It also enables a replacement laptop by renaming the serial-number
 folder on the school server (and updating the registration
 information).

 I am not sure if you are trying to implement this functionality, but
 it is important at a deployment.

 Yours,

 Tony




___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Journal backup on school server

[George Hunt]

Thanks Tony,

I really am enjoying reading your code.  Some tricks there that I'm
learning from.

It seems to me that your retinking of ds-backup makes a wonderful addition
to the ways we can think about ds-backup. I'd like to set up a github repo
which starts from the current laptop.org one, and then branches to all the
different ways ds-backup can be expanded.  Eventually, each branch could
become it's own thing.  It might be nice to incorporate an upgrade feature
in ds-backup, so that when a specific school server is installed, and the
XO's connect to it, the proper ds-backup behavior is substituted.

George


On Sun, Apr 7, 2013 at 7:46 AM, Tony Anderson tony_ander...@usa.net wrote:

 Hi,

 I have attached the ds_backup.py code. Tack on all disclaimers - it has
 not been tested in the XS-0.7 and 12.1.0 environment. It will need some
 revision to deal with the 'shared' XO environment in Lesotho so I will work
 on it while I am there.

 However, it really is not changed from XS-0.7 in areas that matter to you.
 The changes are in the way the Journal objects are handled. Martin Langhoff
 used rsync so that the backup is a snapshot of the local datastore. I am
 trying to establish the server-side as the real Journal with the local
 datastore containing only the currently relevant content. It seems sad that
 kids must delete their Journal because of the limited size of the local
 store. The message: Journal is full is not helpful, because the fact is
 that the store is full, often because of the excessive number of installed
 activities. With a school server, this is totally unnecessary since the
 Sugar Activities are available on the school server so that unused ones can
 be deleted on the XO.

 I had to move to mod_wsgi for Django (library). As far as I know, there is
 no change needed client-side. Apache needs to have wsgi installed (in
 XS-0.7 it is). Then all that is needed is a file in httpd/conf.d.
 I have also attached the one I am using.



 On 04/07/2013 06:04 AM, George Hunt wrote:


 Hi Tony,

 I'm playing with ds-backup now, because one of the dependencies for
 ds-backup is mod_python which was dropped from fedora 18, in favor of
 mod_wsgi. It seems like a great time to rethink the issue.

 I cannot find the working code you mention in this email in my google
 stack.  Can you send it to me again?

 I'm also talking with Tim Moody, about using puppet client as a way of
 pushing and modifying the configuration of XO's in the classroom.  Do
 you have any experience, or ideas, about such a proposal? I'm not sure
 whether ds-backup might be used to introduce puppet client back into XOs.

 George

 On Sun, Dec 9, 2012 at 12:25 PM, Tony Anderson tony_ander...@usa.net
 mailto:tony_ander...@usa.net** wrote:

 Hi,

 I will send you the scripts as soon as I have a chance to try them
 on 12.0.1 and XS-0.7. Last year they were running on xs-0.6 and
 build 852 at Saint Jacob. There shouldn't be too much difference,
 although so far on most of the code I have been wrong by about 3 days!

 The identity problem really isn't that hard. A lot of the mailing
 list discussion seems related to running Sugar on something besides
 an XO.

 The user is identified on the server by the serial-number. For the
 backup, this is not a problem. The problem comes from OLPC where an
 XO is shared. Then there needs to be something like a login to
 identify the actual user. There also needs to be separate Journals
 on the XO and separate backups on the school server.

 I should be able to give you working code within a week (I leave
 Kigali on 12/23 so it must be before then).

 Yours,

 Tony


 On 12/09/2012 04:11 PM, George Hunt wrote:

 Wow Tony,

 Your solution definitely needs to be part of the mix as we go
 forward.
 Can you send me copies of the scripts, or the changes you made to
 accomplish these objectives?  If we are going to look at the
 serialnumber-user-identity issue, we might be making changes to
 the same
 packages.

 Have you worked through the changes needed to add user identity to
 journal backups? As Paul Fox was suggesting, I think there will
 be lots
 of ripple effects, if and when we start adding additional users,
 in
 addition to the user olpc.  Or maybe, someone will come up with
 a
 simplifying assumption, or approach.

 George



 On Sun, Dec 9, 2012 at 4:49 AM, Tony Anderson
 tony_ander...@usa.net mailto:tony_ander...@usa.net
 mailto:tony_ander...@usa.net mailto:tony_ander...@usa.net**
 __

 wrote:

  Hi, Sameer

  I got an email from Nick Doiron re his visit to the Marshall
  Islands. He mentioned that you gave him a script to create
 a csv
  from the Journal backup.

  As I have mentioned several times

[Server-devel] ds-backup

[George Hunt]

Hi Daniel,

I noticed that you were working on ds-backup recently.  I was trying to get
XSCE running on fc18, and encountered the issue of fedora dropping
mod-python (see https://fedorahosted.org/rel-eng/ticket/5165).

Is it ok to just carry the fc17 version along in our repo, and try to see
if it just works, or should we set a goal of rewriting to mod_wsgi?

I don't have much experience in this sort of thing.

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] XSCE wants to become a framework for the next 10 years

[George Hunt]

Tony, et al,

The group of developers, working on the XSCE, are indeed attempting to
build upon the good work that Daniel Drake did on the XS-0.7. But we are
trying to extract the essential information from the history of the school
server up to this point.

   - The XS-0.6, based upon FC4, was released in the 2008 time frame.
   - Nepal, Australia, Uruguay, perhaps for their own and different
   reasons, deviated from this released version 2008-2012.
   - XS-0.7 was released for use in Nicaragua based upon Centos in early
   2012.

Our analysis of this history has been that the monolithic nature of the
punji, anoconda build process is not helpful.  If the functionality of the
school server could be dropped on top of a current fedora build, all of the
hardware specific configuration would be handled by the general Fedora
community -- our school server software doesn't need to change to
accomodate arm, or x64.

But as with any basic restructuring, starting from the ground up, we need
to walk before we can run.  Whether it is reinventing the wheel or not --
networking needs to work flawlessly. We have determined that one the the
hardware platforms we need to support is the XO itself. The XO uses
NetworkManager as it's networking frontend, so to be compatible, we have
needed to learn how to configure NM.  Squid, ejabberd, and iptables need to
play in all configurations of network adapters.

In addition, if we are thinking for the next 10 years, we wanted a more
modular plugin-like structure for adding additional services.

So I believe Tony, you are correct, we seem to be reinventing the wheel.
But it's my hope we are getting this wheel ready to carry a much heavier
weight.  We are hoping that by the third quarter of this year, the XSCE
might be to the point where it is a drop in replacement for XS-0.7. At that
point your good suggestions might be extremely useful.

We are trying to provide a software framework that is attractive and
flexible enough, so that in the future, the next Nepal, Australia, Uruguay
will not feel the need to go their own way.

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Is a Community Edition of XS happening? or should it?

[George Hunt]

Hi all,

I just got the included message from Adam Holt, after his on the ground
experiences installing a school server in Madagascar, and apparently
struggling to get ejabberd working.  It points up a situation which I think
we should think about.

A lot of people I've talked to, think the School Server status quo is not
good enough. It is not meeting the needs of schools, teachers, and
students. Many are beginning to go their own way. The centrifugal force is
building:

   - OLPC Australia is looking to simplify the XS, to include just ejabberd
   for collaboration and have it run on the XO-1.75. Eliminate dhcp in favor
   of avahi, eliminate Moodle, Squid, Named. Sridhar Dhanapalan wants to get
   to the point where the individual teacher in the classroom can set it up.
   Jerry Vonau has been hired to muscle up support for the upcoming deployment
   of 50,000 XO's with one XS in each classroom.
   - In the Philippines, through bad advice, the local technicians started
   trying to use the Australia version of the XS.  They didn't have the local
   sysadmin skills to add back in named and dhcpd, which had been removed for
   Australian deployment. They're looking for a better solution.
   - Adam Holt has been soliciting ideas from the support gang for finding
   a XS solution that just works.
   - Jamaca is making Moodle central to its deployment strategy, but it
   needs some predictability in terms of school server depoyment.
   - Tony Anderson and Abhishek Singh,in the Nepal deployment, have their
   own XS image tailored to their own needs.


But I also think that the support that Boston has given to the XS has been
essential. Daniel Drake's XS-0.7 brought together many of the improvements
that have accumulated over the last few years.

Maybe we're at the point where Red Hat was, when it split the Enterprise
Linux from Fedora Core. EL would have a slower release cycle, and pick up
the features that had been well tested via the six month Fedora release
cycle.

Sridhar seems to have the energy, resources, and management skills to make
the stripped down XO-XS happen.  Tony Anderson, Sameer Verma, Abhishek
Singh have all expressed to me their willingness to contribute to some
joint effort.

From my point of view, the challenge is to keep it simple, and to start
working towards a structure where all of us can take a small piece, work on
it, and contribute it back to the common effort.

George

-- Forwarded message --
From: Adam Holt h...@laptop.org
Date: Sun, Jul 29, 2012 at 8:03 AM
Subject: fix at last? changing XS hostname dilemmas
To: Mitchell Seaton meaton...@gmail.com, Craig A. Perue 
craig.pe...@gmail.com, Sameer Verma sve...@sfsu.edu, George Hunt 
georgejh...@gmail.com, Xavier Carcelle xavier.carce...@gmail.com
Cc: Alex Kleider aklei...@sonic.net, Jerry Vonau jvo...@shaw.ca


Skype excerpt :)

[7:39:35 AM] Jerry Vonau: Sorry I haven't gotten back to you earlier, think
I know what the issue is with ejabberd if you change the hostname.
[7:40:01 AM] Canoe Berry: Really??
[7:40:49 AM] Jerry Vonau: ejabberd creates a pem.cert based on the hostname
when first installed, change the hostname and it becomes invalid
[7:42:54 AM] Jerry Vonau: quick fix is to delete /etc/ejabberd/ejabberd.pem
and re-install ejabberd.
[7:43:40 AM] Jerry Vonau: or create your own cert and alter the config file
to use the new one instead.
[7:45:40 AM] Jerry Vonau: by the way I have ejabberd running on XO-1.75.
[7:48:54 AM] Canoe Berry: Sweet, can I forward to Craig/Sameer for Jamaica?
[7:49:15 AM] Jerry Vonau: as for the dns issue, think the model used is
flawed, you need full control over the dns, think we should be using avahi
on the XS and XO and use .local for the domain.
[7:49:37 AM] Jerry Vonau: have that working too
[7:50:02 AM] Canoe Berry: Are you suggesting Madagascar/Nosy Komba/OLPC
France's use of dsnmasq is doomed, as Mitch Seaton suspected?
[7:51:47 AM] Jerry Vonau: sort of, with the push to ip6 the clients use
ipv6 first to resolve the names and that is where the delay is introduced.
[7:52:11 AM | Edited 7:52:56 AM] Jerry Vonau: when registering

If someone can clean this up and post to server-devel@l.o that's be awesome!
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] armv7hl vs armv7l

[George Hunt]

Hi Peter,

You probably know the answer to this question off the top of your head.

I've played with fedora's Trimslice armv7hl, using it to recompile XS rpms.
Now in conversation with OLPC-Australia, I've agreed to try to apply my
stuff to the XO-1.75 pre-release 12.1.0, which I believe is based upon FC17.

Question: Is my easiest path to basically start over, either building up a
cross compiling tool chain, or maybe try to compile the XS rpms on an
armv7l machine natively, as I did with the TS, (the XO itself seems the
obvious choice).

I had trouble earlier getting a tool chain together to run on FC17, on top
if parallels, on my MAC.

Do you have any advice?

Thanks,

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] armv7hl vs armv7l

[George Hunt]

Thanks Peter,

I was confused when I installed latest 12.1.0 on an XO and issued uname
-a, to see the response come back armv7l, rather than armv7hl.  I was
thinking that yum would be confused by the difference.

I'm glad that the trimslice generated rpms I have will be usable.  I'll
need to learn how to override the default arch, so that yum will do what I
want it to do. But I have google for that!

Thanks for your help,

George

On Fri, Jul 6, 2012 at 6:39 AM, Peter Robinson pbrobin...@gmail.com wrote:

 On Fri, Jul 6, 2012 at 10:42 AM, George Hunt georgejh...@gmail.com
 wrote:
  Hi Peter,
 
  You probably know the answer to this question off the top of your head.
 
  I've played with fedora's Trimslice armv7hl, using it to recompile XS
 rpms.
  Now in conversation with OLPC-Australia, I've agreed to try to apply my
  stuff to the XO-1.75 pre-release 12.1.0, which I believe is based upon
 FC17.

 It is indeed, it's using the F-17 arm hardfp release.

  Question: Is my easiest path to basically start over, either building up
 a
  cross compiling tool chain, or maybe try to compile the XS rpms on an
 armv7l
  machine natively, as I did with the TS, (the XO itself seems the obvious
  choice).

 If you have a trimslice why don't you use that and compile natively?
 In Fedora everything is compiled natively with no cross-compilation.

  I had trouble earlier getting a tool chain together to run on FC17, on
 top
  if parallels, on my MAC.

 To be honest I've never cross compiled any ARM packages.

  Do you have any advice?

 Compile natively :-)

 If you have a Trimslice, Pandaboard or even an XO 1.75 you can compile
 on all of those using the standard distros. On any of the platforms
 you can yum install or yum groupinstall  anything you may need and
 build directly. You might want to add an ext4 formatted usb HDD to use
 as the storage for building on those platforms as they tend to be a
 bit quicker than SD card storage.

 Peter

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] ARM on XS -- how can I integrate my work?

[George Hunt]

Thanks Daniel for the reality check.  I often seem to get caught up in
wanting to learn about something without seeing the larger context.

I was looking at the function per watt, and not seeing the
development/testing/support costs involved in adding another distribution.

In retrospect, I was also ignoring another fact, that the fitpci has a dual
core atom processor that draws 7 watts just like the ARM Trimslice -- ARM
is not really necessary to achieve low power school serving.

All is not lost however, I enjoyed learning about ARM, and packaging rpms
-- skills I may find useful.

George

On Mon, Jun 25, 2012 at 11:27 AM, Daniel Drake d...@laptop.org wrote:

 On Sun, Jun 24, 2012 at 8:14 AM, George Hunt georgejh...@gmail.com
 wrote:
  I'm not done yet, but I've been making progress on porting XS code to
 ARM by
  making modifications to DSD's  XS-0.7.  Upon his suggestion, I have been
  basing my work on the srpms posted at
  http://xs-dev.laptop.org/xsrepos/stable/olpc/xs-0.7/source/.
 
  Now that I've got some of the services running, I'm wondering how to
  contribute to the XS codebase.  What I'd prefer is to contribute deltas
 from
  XS-0.7 that use `uname -p` to enable the appropriate path through the
  startup scripts.

 That kind of approach would suggest supporting both CentOS and F17+.
 I'm not sure if thats the direction we'd want to go - supporting 2
 platforms has its costs. It might be preferred to do a full migration
 to F17. You'll need clarification from this from Martin, who's away
 until next month.

  Earlier, I started using the git sources at dev.laptop.org, and I
 discovered
  that there did not appear to be an obvious set of git repos,
 corresponding
  to XS-0.7. Paths for repos that had the most recent changes included:
 
  /packages/
  /projects/
  /bios-crypto/
  /users/martin/
 
  Any suggestions on how we should proceed?

 I think you just need to sit down and hunt these down. It shouldn't
 take you long, using version numbers found in the most recent RPM
 packages, etc.

 Many of the XS component webpages have links to the git repo, e.g.
 http://wiki.laptop.org/go/XS-rsync

 If there are ones that don't, maybe you can add the links.

 And if you really get stuck, you can ask for help for a particular package.

 Daniel

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] School server on ARM related question: Systemd on fc17 doesn't see ethernet dongles or USB sticks

[George Hunt]

Hi all,

There are no entries in /var/log/messages or dmesg relating to my test USB
stick at /dev/sdb1, even though during the boot process, dracut sees an 8
GB sdb as well as the 132GB hard disk on sda that is the rootfs (so I think
the proper drivers are in the kernel). I'm playing with the trimslice root
fs and kernel snapshots available at
http://fedoraproject.org/wiki/Architectures/ARM.

I also cannot configure an ethernet dongle, which the kernel doesn't appear
to respond to, when inserted.

Any pointers to documentation, or what modules's source I might study? I
don't really understand how HAL was replaced, or why there are no entries
in the kernel log.

Thanks,

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] School server on ARM related question: Systemd on fc17 doesn't see ethernet dongles or USB sticks

[George Hunt]

Thanks,

George

On Sat, May 19, 2012 at 4:06 PM, rihowa...@gmail.com rihowa...@gmail.comwrote:



 On May 19, 2012, at 12:22 PM, George Hunt wrote:

 Hi all,

 There are no entries in /var/log/messages or dmesg relating to my test USB
 stick at /dev/sdb1, even though during the boot process, dracut sees an 8
 GB sdb as well as the 132GB hard disk on sda that is the rootfs (so I think
 the proper drivers are in the kernel). I'm playing with the trimslice root
 fs and kernel snapshots available at
 http://fedoraproject.org/wiki/Architectures/ARM.


 This is a known problem with the TrimsSlice and has been documented on the
 fedora arm list
 Fedora 17 for ARM is just approaching the end of alpha development and a
 fix may appear soon if it has not occurred already.
 The place to ask about F17 for ARM is at the fedora arm email list.

 I also cannot configure an ethernet dongle, which the kernel doesn't
 appear to respond to, when inserted.

 Any pointers to documentation, or what modules's source I might study? I
 don't really understand how HAL was replaced, or why there are no entries
 in the kernel log.

 Thanks,

 George
 ___
 Server-devel mailing list
 Server-devel@lists.laptop.org
 http://lists.laptop.org/listinfo/server-devel



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] xs-otp tarball pkg?

[George Hunt]

Hi all,

I've finally got my arm compiler verified, and booted my first ARM kernel
on an XO (using a Trimslice native compiler and fedora 17).

So now I'm trying to repackage all the XS packages for ARM. I don't know
enough to interpret the cryptic comment in the changelog of the xs-otp git
repository at https://dev.laptop.org/git/users/martin/xs-otp/ - use
fedpkg.  I've googled for a fedora package with xs-otp unsuccessfully.
When git only includes the spec file, is there a conventional place to look
for the source tarball?

For most of the packages, I've taken the easy way out, and used DDrake's
source rpms at http://xs-dev.laptop.org/xsrepos/stable/olpc/xs-0.7/source/as
a basis for my arm rebuilds.

 But I'm wondering what diff's there might be with the git directories,
and/or if I should be trying to reconcile his builds with what's at
https://dev.laptop.org/git/.

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] [OLPC School Server] Comment: First Learnings about ARM

[George Hunt]

 I need some pointers.

My target machine is a Dreamplug.

In the Fedora 17 arm distro, there are separate releases for ARMv5 and
ARMv7.  I have been playing with the ARMv5 download which includes kernels
for tegra, imx, omap, and armv5tel (the latter is what I have been using).
I created a FAT16 partition on an sd card, used uboot's mkimage to package
the kernal for uboot to load. I can execute  the uboot command: usb start;
fatload usb 1 0x640 /boot/uImage

But when I load the uImage to memory at 0x640, and bootm 0x640 (how
do I know that that's the correct load address -- I copied the trimslice
directions), I get no response.  When I do the same with a kernel supplied
by compulab for the dreamplug, it seems to function properly, ie. it boots
to the point where it is looking for the rootfs (it gets further in the
boot process).

Is there any documentation on how to proceed in FC17 with the supplied
kernels? Maybe I should just cross compile to get to a uImage?

George

On Sun, Apr 15, 2012 at 2:28 AM, comment-re...@wordpress.com 
comment-re...@wordpress.com wrote:


 New comment on your post First Learnings about ARM
 Author : kevix (IP: 98.113.63.95 ,
 pool-98-113-63-95.nycmny.fios.verizon.net)
 E-mail : kevin.m...@verizon.net
 URL: http://kevix.myopenid.com
 Whois  : http://whois.arin.net/rest/ip/98.113.63.95
 Comment:
 George, if you are interested, I chat with OLPC folks on a daily basis on
 a private XS. Some of the folks are familiar with ARM. Let me know if you
 want the server info.

 You can see all comments on this post here:

 https://schoolserver.wordpress.com/2012/04/14/first-learnings-about-arm/#comments

 Permalink:
 https://schoolserver.wordpress.com/2012/04/14/first-learnings-about-arm/#comment-39
 Trash it:
 http://schoolserver.wordpress.com/wp-admin/comment.php?action=trashc=39
 Spam it:
 http://schoolserver.wordpress.com/wp-admin/comment.php?action=spamc=39


 You can reply to this comment via email as well, just click the reply
 button in your email client.



___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] XS procurement recommendations (new hardware - Nosy Komba, Madagascar)

[George Hunt]

I like the Fitpc2i, which has two ethernet ports.  I recently purchased one
and would be happy to do whatever testing makes sense. (haven't yet
purchased the sata drive though). I can atest to the 5 watts, idle.  The
blurb that came with it says it's capable of 2 watts standby, but we'd need
to see if it can do wake on lan.

George

On Sun, Apr 15, 2012 at 1:59 PM, Sameer Verma sve...@sfsu.edu wrote:

 On Sun, Apr 15, 2012 at 7:48 AM, Mitchell Seaton msea...@ekindling.org
 wrote:
  Dear all XS Server devel's,
 
  The OLPC France team is looking to provide a new XS server machine for
 their
  deployment in Nosy Komba. The current XS is not functional due to a
 current
  failed disk (cause to be discovered). Myself, and Adam Holt are planning
  work with Xavier (OLPC France), to help support the installation and
 final
  configuration of this new machine during June this year. We thought it's
  important to consult the community here ahead of this purchase decision.
 
  Firstly, we'd like to seek some clear hardware suggestion(s) for a
 potential
  new XS. Solar/batteries are current powering method at the deployment
  site. XO volume to be currently supported is 160, and should allow room
 for
  additional deployments. I believe, the investment should be positioned
 for
  min. 5yrs lifespan, and to cope in local heat, humidity, salt, dust,
  conditions.
 
  We're aware of some hardware that has been working in the field - I've
  particular noted MSi (WindBox) machine has been working successfully in
  Nepal, Haiti and Philippines (George's donation). This is also a
 follow-up,
  in some ways, on George's previous thread although more for deployment
 than
  testing/dev. We're not looking at any XO-1.5/XO-1.75 for form-factor, or
 ARM
  as a platform (know this will come and be more stable in future releases
 to
  come). Low-power is a priority, this will help conserve more power supply
  for XO charging. I assume we are still looking at 32-bit x86 arch and
  potentially preferring Atom processors (unless otherwise advised). Any
  current models in the marketplace for recommendation?
 
  Purchase of requirement machine looks to be possible from France/EU for
  deliver to Madagascar by the deployment/volunteer team (end of May).
  It's crucial to identify the hardware choice relatively soon (during next
  month period). The machine will likely be pre-configured by Xavier prior
 to
  deliver.
 
  Look forward to your comments here.
 
  --
  Mitchell Seaton
  eKindling Technology Support
  msea...@ekindling.org
  www.ekindling.org
 
 
  ___
  Server-devel mailing list
  Server-devel@lists.laptop.org
  http://lists.laptop.org/listinfo/server-devel
 

 We use a SolidLogic box (OLPCorps box) in Jamaica at both locations.
 Its no longer sold (not sure why?) but uses a VIA C7 processor and is
 fanless. Draws 25W at the AC end of the power brick (DC 19V). Here's a
 list of what they offer.
 http://www.logicsupply.com/categories/fanless_systems

 A FitPC2 should work as well. It pulls 15W at the AC end of the power
 brick (DC 12V). The only concern I had with a FitPC2 was the heat.
 They do sell an after market heat sink that helps. If you wall-mount
 the box, you could fashion a chimney on the heatsink to create a
 continuous draft of air (hot air rises and will create an air column
 that pulls in cool air from the outside) to pull the heat away from
 the sink.

 We tested a FitPC2 (See appendix

 http://wiki.laptop.org/images/4/46/Testing_the_OLPC_School_Server_Benjamin_Tran_SFSU.pdf
 )
 and while the tests were aggressive (all 160 XOs will hit the server
 in 60 seconds!), the FitPC2 should be ok.

 http://www.fit-pc.com/web/fit-pc/fit-pc2-models/
 http://www.fit-pc.com/web/fit-pc/accessories/

 cheers,
 Sameer
 --
 Sameer Verma, Ph.D.
 Professor, Information Systems
 San Francisco State University
 http://verma.sfsu.edu/
 http://commons.sfsu.edu/
 http://olpcsf.org/
 http://olpcjamaica.org.jm/

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Looking for new low power server hardware candidate

[George Hunt]

When I took the picture in the following url, I was focusing on what it
would take to run off of 12V deep cycle battery:

http://schoolserver.wordpress.com/xs-installation/xs-0-7-running-on-xo-1-5/

I'm concerned with packaging,  and physical robustness in a real school
setting.  Maybe we could get someone with ME skills to dream up a cheap
package for all the accessory items.

We probably don't need the DC to DC inverter and the usb hub.  But then we
don't have an extra port for sneaker-net, or an adult sized usb keyboard.

At the fall 2011 summit, there was a general call for a turnkey XS that
just worked.  If we could solve the form factor problem, the XO1.75 might
be a good solution.

I think it was Sameer who was telling me that in Australia, they are
thinking about one XS per classroom. In that setting, seems to me that
XO1.75 (even with only 512MB memory) would be more than adequate.

George

On Tue, Apr 10, 2012 at 10:55 PM, John Watlington w...@laptop.org wrote:


 Why not an XO-1.75 ?

 On Apr 10, 2012, at 8:04 PM, Martin Langhoff wrote:

  On Mon, Apr 2, 2012 at 2:03 AM, George Hunt georgejh...@gmail.com
 wrote:
  I left my fitpc2 and msi servers in the Philippines, hoping they would
 be
  pressed into service in a classroom situation.  So now I'm in the
 market for
  another toy.
 
  If you have time towork with us through some hitches, I'd recommend an
  ARM server. At this stage I'd say one of the Marvell/Globalscale
  Plug servers (dreamplug for example), or a trimslice.
 
  Either option will need a combination of the OS on internal SD/eMMC
  and the storage on an ext HDD (via USB probably).
 
  cheers,
 
 
  m
  --
   martin.langh...@gmail.com
   mar...@laptop.org -- Software Architect - OLPC
   - ask interesting questions
   - don't get distracted with shiny stuff  - working code first
   - http://wiki.laptop.org/go/User:Martinlanghoff
  ___
  Server-devel mailing list
  Server-devel@lists.laptop.org
  http://lists.laptop.org/listinfo/server-devel
 


___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

Re: [Server-devel] Looking for new low power server hardware candidate

[George Hunt]

I'd love to be involved in getting an arm server going.  I have a
dreamplug.  Which distro are you thinking would be a good base?

What sort of hitches? If there's discussion in the archives, I'll do
reading to get up to speed.  I might help if I knew when or who.

George

On Tue, Apr 10, 2012 at 8:04 PM, Martin Langhoff
martin.langh...@gmail.comwrote:

 On Mon, Apr 2, 2012 at 2:03 AM, George Hunt georgejh...@gmail.com wrote:
  I left my fitpc2 and msi servers in the Philippines, hoping they would be
  pressed into service in a classroom situation.  So now I'm in the market
 for
  another toy.

 If you have time towork with us through some hitches, I'd recommend an
 ARM server. At this stage I'd say one of the Marvell/Globalscale
 Plug servers (dreamplug for example), or a trimslice.

 Either option will need a combination of the OS on internal SD/eMMC
 and the storage on an ext HDD (via USB probably).

 cheers,


 m
 --
  martin.langh...@gmail.com
  mar...@laptop.org -- Software Architect - OLPC
  - ask interesting questions
  - don't get distracted with shiny stuff  - working code first
  - http://wiki.laptop.org/go/User:Martinlanghoff

___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] Looking for new low power server hardware candidate

[George Hunt]

Hi everyone,

I left my fitpc2 and msi servers in the Philippines, hoping they would be
pressed into service in a classroom situation.  So now I'm in the market
for another toy.

I remember Sameer saying somewhere in the email compost that he had found
new interesting server hardware, but a google search of my email heap was
not successful.

Does anyone have proposals for hardware that fills the XS requirements on
the low power end of the spectrum?

I've blogged about  my school server experiences at
http://schoolserver.wordpress.com.

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] School Server system speed diagnotic

[George Hunt]

Hi all,

In the Philippine deployments, the local decision makers have tried to buy
locally, and the machines use roughly 80 watts, idling, bursting to 120
during startup.

The inadequate power situation causes rolling brown outs and load
shedding.  I've been unable to document the decrease in performance they
might encounter if they shifted their usage to a dual core Atom 30 watt MSI
or a 10 watt fitpc. Selling the idea of deep cycle batteries for school
servers, would be a lot easier if we didn't need so much power drain!

For lack of anything better, I started asking them to time the startup time
until the school server presented a logon prompt.  Yet, I also noticed that
if network interfaces were down, or if there was no wan connectivity, there
would be timeouts which would add noise to such an approach.

As I have understood what I've read, the ejabberd service is currently the
performance bottleneck.  I don't have much familiarity with cpu test
suites.  Does anyone have a suggestion on benchmark programs that would be
useful in predicting actual performance in the schools? If no one has
experience in the area, I can do the googling and reading.  Sometimes it
makes sense to ask first.

Thanks
George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] XS on XO

[George Hunt]

Hi Everyone,

Following suggestions on this list, and a cookbook at
http://wiki.laptop.org/go/Ubuntu_On_OLPC_XO, I loaded kernel 2.6.31 from
olpc build 860 on top of XS-0.7.

After a few false starts, it comes up!

Now I'm confronted with the same problem I was having on the FitPc I was
trying to get going with XS-0.7.

The details:

   - Downloaded XS-0.7 and verified md5sum
   - Tried to load XS-0.7 onto FitPc2 from USB stick created by FC16.
   Failed at that.
   - Shifted over to CD install, and that worked.
   - Did the xs-setup domain name, and xs-setup-network, and got
   networking going (gateway, and ping by number to internet).
   - Started looking into failure of named (many errors in
   /var/log/messages)
   - Tried the Centos stock named.conf -- caching works if resolv.conf
   changed to include nameserver localhost as indicated by dig yahoo.com
   (second named querry served out of cache in much less time)

So I don't know enough to start looking into why dig yahoo.com would work
and ping yahoo.com would fail.

Any ideas where to look?

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel

[Server-devel] named problem resolved-but not understood

[George Hunt]

Hi again,

Please disregard my previous question. . .

Trial and error indicates that if the two option elements query-source
port 53 and query-source-v6 port 53 are commented out, caching, ping,
and default contents of resolv.conf function according to plan.

Why did it work before? Maybe CentOS does something different?

George
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel