Re: Looks like apt-get autoremove Removed Too Much.

[Martin McCormick]

David Wright  writes:
> Pleasure. Nice to know someone else does that too.

I am happy to report that the recording ended so I
rebooted with fingers crossed. The messages from the boot
confirmed that network-manager started and reported okay and the
interface came up with it's assigned IP address exactly as it is
supposed to do. 

If anything else turns up missing, I know what to look for, now.
Thank you.

Martin

Re: Looks like apt-get autoremove Removed Too Much.

[David Wright]

On Tue 11 Apr 2017 at 15:58:26 (-0500), Martin McCormick wrote:
> David Wright  writes:
> > /var/log/apt/history.log should contain something like
> > 
> > Start-Date: 2017-03-02  06:48:56
> > Commandline: apt-get autoremove
> > 
> > followed by the packages removed. If, during the removal, some service
> > was stopped, or whatever, you might see some output somewhat buried in
> > /var/log/apt/term.log at the corresponding time.
> 
>   I am rather sure I found it thanks to your suggestion.
> Unfortunately, the system is running a cron job to record a radio
> program right now so I can't cause a reboot without ruining the
> on-going recording

Pleasure. Nice to know someone else does that too.

Cheers,
David.

Re: Looks like apt-get autoremove Removed Too Much.

[Martin McCormick]

David Wright  writes:
> /var/log/apt/history.log should contain something like
> 
> Start-Date: 2017-03-02  06:48:56
> Commandline: apt-get autoremove
> 
> followed by the packages removed. If, during the removal, some service
> was stopped, or whatever, you might see some output somewhat buried in
> /var/log/apt/term.log at the corresponding time.

I am rather sure I found it thanks to your suggestion.
Unfortunately, the system is running a cron job to record a radio
program right now so I can't cause a reboot without ruining the
on-going recording but I did as you suggested and found the file
as described. in amongst the no-longer-needed packages in the
list of removed software was network-manager, one that is quite
needed on this system. Since a network connection was already
established, nothing stopped at the time network-manager was
removed but the system couldn't then setup eth0 when it came up
again after the reboot I gave it to see if all was well. It
definitely was not well, having no eth0 on reboot.

I had to manually issue the following commands on eth0 to
make the system able to use apt-get install network-manager:

#Bring eth0 up.
ifconfig eth0 inet 192.168.1.81 netmask "255.255.255.0"

#Setup a route to our internet router.
route add 192.168.1.1 eth0

#Let the system know where to route packets to the big wide world.
route add default gw 192.168.1.1

I forgot that last step at first and apt-get install network-manager 
would crash immediately because the system had no routing
information outside of our private network.  It couldn't do DNS
lookups nor could it reach the repositories. After I got that
done, it installed all the packages so I expect it to do dhcp
next time I reboot.

Again, my thanks to everybody who helped.

Martin

[COLABORACAO] Kexec - Reiniciando o Linux sem reiniciar o Hardware

[Henrique Fagundes]

Prezados colegas,

Gostaria de colaborar com a comunidade com essa dica!

Segue o link:
https://www.aprendendolinux.com/kexec-reiniciando-o-linux-sem-reiniciar-o-hardware

Espero que seja útil para alguém.

Atenciosamente,

Henrique Fagundes
henri...@linuxadmin.com.br
Skype: magnata-br-rj
Linux User: 475399

http://www.aprendendolinux.com/
http://www.facebook.com/PortalAprendendoLinux
http://youtube.com/aprendendolinux/
http://twitter.com/aprendendolinux/
__
Participe do Grupo Aprendendo Linux
https://groups.google.com/forum/#!forum/portal-aprendendo-linux

Ou envie um e-mail para:
portal-aprendendo-linux+subscr...@googlegroups.com

Re: oom-killer no funciona como debería.

[Santiago Vila]

On Mon, Apr 10, 2017 at 11:31:59AM +0200, luis gil wrote:

> ¿Como puedo saber si mis aplicaciones hacen overcommit en proporción 1:2?

Lo que dije sobre 1:2 era solamente un ejemplo.

Mientras las aplicaciones hagan overcommit, se debe poner swap para compensar.

> Mi RAM es de 32 G. Me parece excesivo una swap de ese tamaño.

Todo depende de las aplicaciones que estés usando. Si tus aplicaciones
"piden" 64 GB de memoria pero solamente usan 32 GB "normalmente", más
vale que tengas esos 32 GB adicionales en swap para el día en que
realmente quieran usar la memoria que han pedido.

> Según había leído por ahí en varios sitios (no he encontrado mucha
> información sobre este tema), cuando pones el overcommit_memory a 2, se
> fija un nivel de memoria RAM a partir del cual se empiezan a matar
> procesos.

No es exacto. Se fija un nivel de memoria comprometida *total* a
partir del cual se empiezan a matar procesos.

Si este nivel es inferior a la cantidad de swap más la cantidad de
RAM, entonces los procesos morirán antes de que comprometas toda la
memoria disponible (swap + RAM) lo cual es un desperdicio.

Por este motivo el valor del porcentaje, para que tenga sentido,
debería ser >= 100.

Estamos hablando de la memoria "comprometida", esa que a menudo supera
con creces la memoria realmente utilizada.

> Este nivel se calcula realizando la siguiente operación Cantidad
> SWAT + Cantidad RAM * ratio / 100. Poniendo el ratio a 75 el nivel se queda
> un poco por debajo de la RAM, de esta forma se consigue que el sistema no
> empiece a swapear.

Tampoco es exacto. Para que el swap se utilice lo menos posible
hay otro parámetro llamado swappiness.

Lo que quieres no es que no utilice el swap, sino que los procesos no
mueran por falta de memoria "total".

> [...]

De nada sirve monitorizar la memoria usada por Apache cada 5 minutos
si un atacante externo puede lanzar en dos o tres minutos tantas
peticiones http que el sistema se quede sin memoria entre una
observación y la siguiente.

En mi caso, empecé a interesarme por este tema el día que alguien
lanzó 2000 peticiones http a páginas PHP inexistentes en menos de tres
minutos a un servidor que tenía con Apache en modo prefork.
El servidor hizo "cataplof".

Debería ser imposible, por *diseño* (no solamente tomando mediciones
experimentales y monitorizaciones varias), que la memoria necesaria
pueda crecer indefinidamente porque así lo provoque algo externo.

Por eso recomiendo PHP-FPM + nginx.


Por cierto, el incidente de United Airlines de ayer me ha recordado
esta analogía de Andries Brouwer sobre lo absurdo que es el OOM killer:

https://lwn.net/Articles/104179/

A veces la realidad supera la ficción.

Re: Looks like apt-get autoremove Removed Too Much.

[Brian]

On Tue 11 Apr 2017 at 13:12:58 -0500, Martin McCormick wrote:

>   After using apt-get autoremove on a jessie installation,
> the system is mostly working but it now fails to initialize eth0
> so it comes up with no network address. I looked back through the
> 8 syslog files it had and, of course, there doesn't happen to be
> a syslog taken of a boot or I might be able to see what happened
> when avahi-daemon got the IP address for eth0.

autoremove will only remove packages which are not depended on by other
packages. Its list to autoremove depends on previous history - what was
purged or removed earlier. Only you know what that was.
 
>   For now, I manually setup eth0 by using ifconfig so it
> presently has an IP address and seems to otherwise be working but
> I am not sure what got removed that shouldn't have been.

autoremove only removes what *should* be removed because it is no longer
needed. You agreed to its suggestions.

>   Here is what syslog said when I rebooted the system after
> the autoremove:
> 
> a0 avahi-daemon[1668]: Leaving mDNS multicast group on interface eth0.IPv6 
> with address fe80::211:11ff:fe0e:4e18.
> a0 avahi-daemon[1668]: Leaving mDNS multicast group on interface eth0.IPv4 
> with address 192.168.1.81.
> a0 avahi-daemon[1668]: avahi-daemon 0.6.31 exiting.
> 
> Here it is coming back up.
> 
> a0 avahi-daemon[]: Found user 'avahi' (UID 106) and group 'avahi' (GID 113).
> a0 avahi-daemon[]: Successfully dropped root privileges.
> a0 avahi-daemon[]: avahi-daemon 0.6.31 starting up.
> a0 avahi-daemon[]: Successfully called chroot().
> a0 avahi-daemon[]: Successfully dropped remaining capabilities.
> a0 avahi-daemon[]: Loading service file /services/udisks.service.
> a0 avahi-daemon[]: SO_REUSEPORT failed: Protocol not available
> a0 avahi-daemon[]: SO_REUSEPORT failed: Protocol not available
> 
> Tat's probably where something got zapped that shouldn't have.

This is about avahi, a protocol used on the local network. Nothing to do
with the internet.

> This system configures it's interface via dhcp and that is what is
> now not happening.
> 
>   ifconfig -a shows eth0 all right but it is not set to
> anything until I manually configure it. /etc/network/interfaces
> is okay. There are no huge squawks in the syslog on bootup but 
> sftpd isn't starting, probably because there is no interface to
> listen on. After setting eth0, vsftpd started.

What do you mean by "manually"? How was eth0 configured before?
Why is /e/n/i sound?

> Any constructive suggestions are greatly appreciated as to what
> to look for that should be there but is not.

/var/log/apt.

-- 
Brian.

Re: Looks like apt-get autoremove Removed Too Much.

[David Wright]

On Tue 11 Apr 2017 at 13:12:58 (-0500), Martin McCormick wrote:
>   After using apt-get autoremove on a jessie installation,
> the system is mostly working but it now fails to initialize eth0
> so it comes up with no network address. I looked back through the
> 8 syslog files it had and, of course, there doesn't happen to be
> a syslog taken of a boot or I might be able to see what happened
> when avahi-daemon got the IP address for eth0.
> 
>   For now, I manually setup eth0 by using ifconfig so it
> presently has an IP address and seems to otherwise be working but
> I am not sure what got removed that shouldn't have been.
[...]
> 
> Tat's probably where something got zapped that shouldn't have.
> 
> This system configures it's interface via dhcp and that is what is
> now not happening.
> 
>   ifconfig -a shows eth0 all right but it is not set to
> anything until I manually configure it. /etc/network/interfaces
> is okay. There are no huge squawks in the syslog on bootup but 
> sftpd isn't starting, probably because there is no interface to
> listen on. After setting eth0, vsftpd started.
> 
> Any constructive suggestions are greatly appreciated as to what
> to look for that should be there but is not.

/var/log/apt/history.log should contain something like

Start-Date: 2017-03-02  06:48:56
Commandline: apt-get autoremove

followed by the packages removed. If, during the removal, some service
was stopped, or whatever, you might see some output somewhat buried in
/var/log/apt/term.log at the corresponding time.

Cheers,
David.

Re: Old 32bit PC 650kRam less VidMem 1024x768 will not run on Stretch ok on Jessie

[David Wright]

On Tue 11 Apr 2017 at 12:39:00 (+), GiaThnYgeia wrote:
> 
> 
> Felix Miata:
> > My opinion is that people who want ancient hardware to continue to be
> > supported for the longest possible period of time must participate in
> > the development phase by testing on the very hardware that they want to
> > stay working when it's hardware that developers either no longer have or
> > no longer have motivation to use, and they must do so throughout
> > development, not only in the final weeks before release is expected.
> 
> I'll stick to the "people who want ancient hardware" and ask whether you
> perceive those people as having a choice to "want ancient" hardware or
> whether this is "all" they have.

I would opine, from what I read, that if you want to run mainstream
linux on ancient hardware, then Debian is a pretty good choice to
make. If Debian doesn't support it because it lacks some feature now
seen as essential, then I think you have to look to a specialist
distribution/project run by people who just see these things as a
challenge rather than as systems to use for productive work.

> Do you anticipate those same people to
> be able to start their computing career in developing systems?

No idea what that means.

> Which relates to that world do we want Debian to prevail.  The "free"
> world or the "non-free" world in which we live in?

No point in using the word "free" without saying which meaning you're using.

> > When a problem is found, it needs to be timely reported according to
> > distro policy in the proper place and manner. Reports in the debian-user
> > mailing list are insufficient to fulfill this purpose.
> 
> By who?  By the person who just lost all graphical access to the system?
> By a bug-reporting system that I have yet to see ever working?
> Is a person installing debian expected to know how to access and
> communicate on the command screen of the -recovery mode?

Yes, it's clever that...sort of Catch 22. That's why car breakdown
companies invented Home-Start services, because you can't get to the
garage.

> > You, not we. I'm not here to discuss any policy other than whether this
> > thread's content is suitably located, and with this post I'm finished
> > with this sub-thread.
> 
> Me too ... dealing with such responses
> 
> > Did including iomem=relaxed on your cmdline solve your problem, or did
> > it not?
> 
> I wouldn't know, the person for which I installed Debian for will not
> dare switch to Stretch after this experience, will not even talk about
> it.

So you're not the owner/user of this 2.33Mhz 650kRam PC but are
supporting those who are? I doubt the wisdom of that. But it does
explain why you also want a DE. By the time you add a browser and some
modern web pages, I wonder what sort of performance they'll get.

I got the impression from "What do you really expect the X0org.log to
look like and how helpful would this be to someone with a single
machine?" that you were afraid of breaking the system (in which case,
why are you desparate to upgrade to stretch?). But then I looked
further back and found that originally "it was just an exercise to fix
and backup old data from a WinXP that had become a mesh".

So I can't understand why you ask a question here and, when an answer
is given, just refuse to try it, invent strange reasons to justify not
trying it, and then rubbish Debian and the people here who try to help.

> Did you read about the person who used zsh as a login shell and
> upgraded Jessie to testing and got locked out of his system because zsh
> vanished?  Good thing he knew how to deal with that crisis.

No, but when everything depends on one machine, you don't do stupid
things like upgrading to testing.

> > If yes, say so.
> 
> Are you talking to me?  I am one of this privileged well off people to
> have hardware that can run Stretch and Sid and all kinds of things.
> Send some snail-mail to the less privileged who have lost access to the
> net to ask them.
> 
> > Supporting ancient hardware forever is not going to happen.
> 
> A 2.33Mhz Celeron pc is by no means ancient compared to the Debian claims.

Debian is fulfilling its part of the bargain. It writes "Debian
GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent
permitted by applicable law." People here support others out of the
goodness of their hearts, but they are human. It you demand support
tailored to yourself, then go and pay for it. There's no encumbrance
from Debian; there might be some from the hardware and firmware
manufacturers and you'll have to deal with that yourself.

If you choose to go cheap by using ancient hardware, there's a
trade-off: time and knowledge. If you run a vintage car, you'd better
know how to fix it, or employ a chauffeur. If you drive one, you'd
better know how to double de-clutch, adjust your mixture, advance the
ignition, and so on. And, of course, you help others along the way.

Cheers,
David.

Re: Looks like apt-get autoremove Removed Too Much.

[Hans]

Hi Martin, 

can you get, what packages where deinstalled by autoremove? Should be 
somewhere in the logs.

Best

Hans

lircd is spamming syslog

[Hans]

Hi folks, 

since some days lircd is spamming my syslog with the following message:

Apr 11 20:28:22 localhost lircd-0.9.4c[8159]: Error: Cannot glob /sys/class/
rc/rc0/input[0-9]*/event[0-9]*
Apr 11 20:28:23 localhost lircd[8159]: lircd-0.9.4c[8159]: Error: Cannot glob 
/sys/class/rc/rc0/input[0-9]*/event[0-9]*
Apr 11 20:28:23 localhost lircd-0.9.4c[8159]: Error: Cannot glob /sys/class/
rc/rc0/input[0-9]*/event[0-9]*
Apr 11 20:28:24 localhost lircd[8159]: lircd-0.9.4c[8159]: Error: Cannot glob 
/sys/class/rc/rc0/input[0-9]*/event[0-9]*

At the moment I deinstalled lirc, as my netbook has no infrared diode built 
in. But it looks like a bug, because this is after the last update.

Thanks for any hints.

Best 

Hans

Looks like apt-get autoremove Removed Too Much.

[Martin McCormick]

After using apt-get autoremove on a jessie installation,
the system is mostly working but it now fails to initialize eth0
so it comes up with no network address. I looked back through the
8 syslog files it had and, of course, there doesn't happen to be
a syslog taken of a boot or I might be able to see what happened
when avahi-daemon got the IP address for eth0.

For now, I manually setup eth0 by using ifconfig so it
presently has an IP address and seems to otherwise be working but
I am not sure what got removed that shouldn't have been.

Here is what syslog said when I rebooted the system after
the autoremove:

a0 avahi-daemon[1668]: Leaving mDNS multicast group on interface eth0.IPv6 with 
address fe80::211:11ff:fe0e:4e18.
a0 avahi-daemon[1668]: Leaving mDNS multicast group on interface eth0.IPv4 with 
address 192.168.1.81.
a0 avahi-daemon[1668]: avahi-daemon 0.6.31 exiting.

Here it is coming back up.

a0 avahi-daemon[]: Found user 'avahi' (UID 106) and group 'avahi' (GID 113).
a0 avahi-daemon[]: Successfully dropped root privileges.
a0 avahi-daemon[]: avahi-daemon 0.6.31 starting up.
a0 avahi-daemon[]: Successfully called chroot().
a0 avahi-daemon[]: Successfully dropped remaining capabilities.
a0 avahi-daemon[]: Loading service file /services/udisks.service.
a0 avahi-daemon[]: SO_REUSEPORT failed: Protocol not available
a0 avahi-daemon[]: SO_REUSEPORT failed: Protocol not available


Tat's probably where something got zapped that shouldn't have.

This system configures it's interface via dhcp and that is what is
now not happening.

ifconfig -a shows eth0 all right but it is not set to
anything until I manually configure it. /etc/network/interfaces
is okay. There are no huge squawks in the syslog on bootup but 
sftpd isn't starting, probably because there is no interface to
listen on. After setting eth0, vsftpd started.

Any constructive suggestions are greatly appreciated as to what
to look for that should be there but is not.

Martin McCormick

Re: Old 32bit PC 650kRam less VidMem 1024x768 will not run on Stretch ok on Jessie

[Felix Miata]

GiaThnYgeia composed on 2017-04-11 08:39 (UTC-0400):


Felix Miata:



Did including iomem=relaxed on your cmdline solve your problem, or did
it not?



I wouldn't know, the person for which I installed Debian for will not
dare switch to Stretch after this experience, will not even talk about
it.



If yes, say so.



Are you talking to me?


Who else would I be writing to? You started the thread:
https://lists.debian.org/debian-user/2017/04/msg00091.html
Presumably, you were the "I" with the Prosavage8 S3 Graphics problem about which 
you wrote.



Supporting ancient hardware forever is not going to happen.



A 2.33Mhz Celeron pc is by no means ancient compared to the Debian claims.


While that CPU may well be considered non-ancient, Debian dropped the ball when 
it integrated upstream's KMS without commensurate resources necessary to adapt 
and maintain drivers for all gfxchips other than those from AMD/ATI, Intel and 
NVidia. Support for the ~18 year old Prosavage8 S3 Graphics gfxchip that is the 
reason for this thread's inception has obviously become inadequate, as it has 
for a non-trivial number of other gfxchips designed 15-20 years ago.

--
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)

 Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

Re: Re: how to force mysql to use the name of my machine instead of localhost

[deloptes]

Abdelkader Belahcene wrote:

> thanks a lot for any idea

This has nothing to do with mysql or debian, but with wordpress. Please
refer the manual or subscribe to the wordpress list.
Perhaps someone else here may help you, but it would be OT IMO.

regards

Re: how to force mysql to use the name of my machine instead of localhost

[Richard Owlett]

On 04/11/2017 07:45 AM, Abdelkader Belahcene wrote:

Thanks for reply;
I give more details.
I install wordpress on my machine, it works fine on local,  I mean i if
write in the browser localhost/wordpress, it 's OK.

I can use myIP/worpress on a remote  machine to connect my server, It seems
working  fine, I get the main page,
BUT, CSS files are not read, I know that by prospecting the code source of
the page.

Wordpress store everything in database,  and so the name of the CSS file
for example, unfortunatly,  the path  used is full address of the file like
localhost/wordpress/blabla/style.css
here is the Big Problem

When the connection is local, no problem, since localhost is the host of
the server;

when the connection is  like :   http://myURL/wordpress,   we SHOULD FIND
the CSS file with  http://myURL/wordpress/blabla/style.css



 To solve the problem

   - either  we remove the absolute path for the files, and keep only  the
   relative part,  instead of storing  localhost/wordpress/blabla/style.css, we
   keep only wordpress/blabla/style.css
   - or we replace the localhost by the name of the machine or IP, that is
   what I want to do, I mean instead of storing in DB, localhost I tore the
   name of my machine like bela.enst.dz, the connetion w'll be done with
   this name and the problem is then solved

I used xampp to install apache, mysql and php. But the problem is the same
when I installed each element alone ( apache2, php, mysql)


when I used the command : localhost -f , I got the correct name
tmp$ hostname -f
bela.enst.dz

thanks a lot for any idea

Regards

bela



It depends to a degree on the details of how you installed WordPress 
[that is using apt-get/Synaptic/etc or downloading zip file from 
wordpress.org].
For that I suggest searching the Subject line of this group for about 
the last 5 years.


Some links that this Debian & WordPress newbie has found enlightening:
https://wiki.debian.org/WordPress

If apt-get/Synaptic/etc used for installation
http://localhost ["Apache2 Debian Default Page"]
/usr/share/doc/wordpress/README.Debian

HTH

Re: If Linux Is About Choice, Why Then ...

[Michael Fothergill]

On 9 April 2017 at 21:15, Patrick Bartek  wrote:

> On Sun, 9 Apr 2017 16:25:57 +0100 Michael Fothergill
>  wrote:
>
> > On 7 April 2017 at 19:27, David Niklas  wrote:
> >
> > > On  Mon, 13 Mar 2017 12:30:11 -0700
> > > Patrick Bartek  wrote:
> > > > The Linux mantra has always been "choice," plethoras of choices.
> > > > So why at install time, is there no choice for the init system?
> > > > You get what the developers decide. Yes, you can install a new
> > > > one -- I've done it and it works -- but only after the install.
> > > > It'd be a lot easier, if there were a choice to begin with just
> > > > like whether you want a GUI and which one.
> > > >
> > > > Now, I know with LFS, you get to choose everything, etc.  But is a
> > > > choice of init at install time so outrageous that no one ever
> > > > considered it or is it technically unfeasible or something else.
> > > >
> > > > Just curious.
> > > >
> > >
> > > Because this reply is so late I'm CC'ing you off list.
> > >
> > > I sympathize, I run Gentoo Linux and us OpenRC. I plan on running
> > > Devuan, a Debain derivative that supports lots of different init
> > > systems. Why no one looks at their project and sees the people
> > > involved when making a statistic up for the amount of dissatisfied
> > > systemd users I don't know.
> > >
> > > Sincerely,
> > > David
> > >
> > >
> > ​I have been reading through some of this stuff and I think that the
> > debian users who are fans of the sysinit boot up scripts should
> > switch to running Gentoo.
> >
> > I use Gentoo with the openrc option.
>
> Gentoo is a rolling release.  I prefer the "stable" philosophy of Debian
> -- basically only bug and security fixes. I've been running Wheezy
> now for 5 years, and it's, for all practical purposes, the "same" as
> when I installed it.  After such a time, a rolling release would be a
> completely different animal versionwise.
>

​I see what you mean.  But I would say that once you get Gentoo set up you
can choose from the stable install, a testing type install version and a
more sid like version.

I use the stable option.  Furthermore, after a bit of work you end up with
a bang up to date kernel which updates and reuses the kernel config file
when you do a sync update to the portage tree with a newer stable kernel
from time to time as they are released.

In reality you do a sync update every three weeks and make sure you fix any
broken packages if you encounter any.  The community are good at helping
you with that problem.  That is the equivalent of the bug fixes and
security updates in debian.

If you use the binary package option you can create a set of binaries you
can use in other computers you have locally to maintain etc., assuming they
are using the same architecture ie compile once on the master box and run
everywhere on the other machines ie Gentoo on the master box and "DIY
debian" (sort of) on the clone boxes.

At least I think that's how it works if I understand it correctly.  I am
not sure, but maybe you could get away with not updating the binaries in
the clone boxes too often then that might create some extra stability and
low maintenance if you put some thought into it.

Just a thought.

Regards

MF
​



>
> I've tried rolling releases before.  They are usually cutting edge and
> more problematical (Unless they've gotten a lot better).  I want
> something that works for years and doesn't break. That's one of the
> reasons I chose Debian five years ago. Now, because of the systemd
> thing, I'm looking at alternatives.
>
> > Those who are OK with systemd should stick with Debian.
>
> After much reading, I consider systemd more suited to large, busy
> servers than a desktop box or notebook with just one user.  It's
> like being forced to use a huge tractor-trailer rig with lots of chrome
> and lights and 24 gears when a simple mini-van will do. ;-)
>
> B
>
>

Re: (deb-cat) Chroot des d'Android

[Narcis Garcia]

Apreciat públic, ja he trobat un pedaç per completar el meu propòsit:
Editar /etc/mysql/my.cnf i reemplaçar "bind-address 127.0.0.1"
per "skip-networking".

Llàstima de systemd que no està preparat per aquestes situacions. Espero
que en properes versions de Debian segueixi havent la possibilitat de
prescindir-ne quan sigui necessari.


__
I'm using this express-made address because personal addresses aren't
masked enough at this list's archives. Mailing lists service
administrator should fix this.
El 11/04/17 a les 16:45, Narcis Garcia ha escrit:
> He resolt el problema de SELinux (per Debian 7 i 8):
> # Abans de chroot (Android):
> $ setenforce 0
> # Dins el contenidor chroot:
> $ echo /sys/fs/selinux /sys/fs/selinux none bind 0 0 >>/etc/fstab
> $ mount -o remount,ro /sys/fs/selinux
> 
> I canviant systemd per SystemV a Debian 8 ja puc fer totes les coses que
> puc fer a Debian 7, com per exemple instal·lar i utilitzar
> openssh-server i apache2 com a serveis normals.
> 
> Però estic trabat en el mateix problema amb ambdues versions:
> instal·lant mariadb-server|mysql-server segons el cas, el servei no
> aconsegueix arrencar i no enregistra cap més detall que:
> [FAIL] Starting MariaDB database server: mysqld . . . . . . . . . . . .
> . . . . . . . . . . . . . . . . . . failed!
> 
> He provat de fer-ho amb el servei «rsyslog» en marxa, i tampoc no
> enregistra més informació.
> 
> 
> __
> I'm using this express-made address because personal addresses aren't
> masked enough at this list's archives. Mailing lists service
> administrator should fix this.
> El 11/04/17 a les 10:36, Narcis Garcia ha escrit:
>> He provat a canviar systemd per SystemV al contenidor:
>> $ apt-get install sysvinit-core sysvinit-utils
>> $ cp /usr/share/sysvinit/inittab /etc/
>> $ apt-get remove --purge --auto-remove systemd
>>
>> Però segueixo tinc problemes quan intento instal·lar altres paquets:
>> dpkg (subprocess): cannot set security execution context for maintainer
>> script: Invalid argument
>>
>> he llegit que això té a veure amb SELinux, i n'he instal·lat les
>> utilitats per provar de deshabilitar-lo:
>> $ apt-get install selinux-basics auditd selinux-utils
>> $ setenforce 0
>> no sembla tenir cap efecte potser perquè el nucli Linux d'Android ja ha
>> arrencat amb això;
>> $ selinuxenabled && echo enabled || echo disabled
>> enabled
>>
>> NOTA: Ja he vist les aplicacions «Debian Kit» i «Lil' Debi» de
>> f-droid.org, però el problema és que treballen amb una imatge de disc
>> virtual, i per aquest fitxer cal menjar-se bastant espai
>> d'emmagatzematge. Jo prefereixo compartir l'espai lliure amb la resta
>> d'entorn.
>>
>>
>> __
>> I'm using this express-made address because personal addresses aren't
>> masked enough at this list's archives. Mailing lists service
>> administrator should fix this.
>> El 10/04/17 a les 14:16, Narcis Garcia ha escrit:
>>> Estic preparant una tauleta per ensenyar una aplicació web sense
>>> necessitat d'accés a internet.
>>>
>>> Per això he alliberat els permisos d'Android (root) i hi he instal·lat
>>> una Debian 8 dins un directori amb debootstrap.
>>> Per iniciar-hi serveis en un entorn així, sempre n'havia tingut prou amb
>>> executar-los cuidant una mica les variables d'entorn i punts de muntatge
>>> (fins i tot servia per iniciar escriptoris).
>>>
>>> No sé si és gràcies al venerable systemd que el mateix Bash em dóna
>>> errors com «cannot set terminal process group» i que no puc ni
>>> instal·lar paquets com ssh.
>>> Se m'acudeixen diferents alternatives:
>>> - Tornar a Debian 7
>>> - Canviar systemd per System-V (si és que es pot)
>>> - Que realment hi hagi manera de fer córrer els serveis com abans.
>>>
>>> Algú em pot aconsellar?
>>>
>>
> 

Re: If Linux Is About Choice, Why Then ...

[Richard Owlett]

On 04/11/2017 08:16 AM, Lisi Reisz wrote:

On Tuesday 11 April 2017 13:54:50 Richard Owlett wrote:

'They' never told us, owners of single user laptops, why we
should chose it.


Your quoting lost important context ;<
You quoted only one sentence (as Ric did) of my post of April 8.
But you link to my reply earlier today to Ric's post.
As you argued in another thread, context can be everything,
"Which is referenced here by its own name - Exe GNU-Linux.  When
I first saw it I wondered why a Linux distro was named after a
Windows binary.  The answer is, it isn't.  It is named after the
River Exe in Devon, England."



Because they don't care whether you chose it or not?  Debian offers
alternatives, but had to chose a default.  Other distros have made
their choice.
 Ubuntu has chosen - of course - to go where the money is.

No-one cares what Richard Owlett - or Lisi Reisz - has chosen for his
(her) laptop.  And why should anyone?

Lisi



As to your question, the BMOC (Big Monopolists Of Consumerism) have 
blinded themselves to the potential size of a market.


I get a chuckle following an Android oriented USENET group. There is 
much effort going into how to 'root' an Android object because 
purchaser's actually wish to use those objects for their (the 
purchasers) purpose. After all, Android was touted as a Linux for such 
objects.


The current crop of dinosaurs (MS, Google, Canonical) should remember 
"Those who don't know history are doomed to repeat it.”

[https://www.goodreads.com/quotes/tag/doomed-to-repeat-it]
Seen any T. Rex lately? They were once top of food chain.

Re: (deb-cat) Chroot des d'Android

[Narcis Garcia]

He resolt el problema de SELinux (per Debian 7 i 8):
# Abans de chroot (Android):
$ setenforce 0
# Dins el contenidor chroot:
$ echo /sys/fs/selinux /sys/fs/selinux none bind 0 0 >>/etc/fstab
$ mount -o remount,ro /sys/fs/selinux

I canviant systemd per SystemV a Debian 8 ja puc fer totes les coses que
puc fer a Debian 7, com per exemple instal·lar i utilitzar
openssh-server i apache2 com a serveis normals.

Però estic trabat en el mateix problema amb ambdues versions:
instal·lant mariadb-server|mysql-server segons el cas, el servei no
aconsegueix arrencar i no enregistra cap més detall que:
[FAIL] Starting MariaDB database server: mysqld . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . failed!

He provat de fer-ho amb el servei «rsyslog» en marxa, i tampoc no
enregistra més informació.


__
I'm using this express-made address because personal addresses aren't
masked enough at this list's archives. Mailing lists service
administrator should fix this.
El 11/04/17 a les 10:36, Narcis Garcia ha escrit:
> He provat a canviar systemd per SystemV al contenidor:
> $ apt-get install sysvinit-core sysvinit-utils
> $ cp /usr/share/sysvinit/inittab /etc/
> $ apt-get remove --purge --auto-remove systemd
> 
> Però segueixo tinc problemes quan intento instal·lar altres paquets:
> dpkg (subprocess): cannot set security execution context for maintainer
> script: Invalid argument
> 
> he llegit que això té a veure amb SELinux, i n'he instal·lat les
> utilitats per provar de deshabilitar-lo:
> $ apt-get install selinux-basics auditd selinux-utils
> $ setenforce 0
> no sembla tenir cap efecte potser perquè el nucli Linux d'Android ja ha
> arrencat amb això;
> $ selinuxenabled && echo enabled || echo disabled
> enabled
> 
> NOTA: Ja he vist les aplicacions «Debian Kit» i «Lil' Debi» de
> f-droid.org, però el problema és que treballen amb una imatge de disc
> virtual, i per aquest fitxer cal menjar-se bastant espai
> d'emmagatzematge. Jo prefereixo compartir l'espai lliure amb la resta
> d'entorn.
> 
> 
> __
> I'm using this express-made address because personal addresses aren't
> masked enough at this list's archives. Mailing lists service
> administrator should fix this.
> El 10/04/17 a les 14:16, Narcis Garcia ha escrit:
>> Estic preparant una tauleta per ensenyar una aplicació web sense
>> necessitat d'accés a internet.
>>
>> Per això he alliberat els permisos d'Android (root) i hi he instal·lat
>> una Debian 8 dins un directori amb debootstrap.
>> Per iniciar-hi serveis en un entorn així, sempre n'havia tingut prou amb
>> executar-los cuidant una mica les variables d'entorn i punts de muntatge
>> (fins i tot servia per iniciar escriptoris).
>>
>> No sé si és gràcies al venerable systemd que el mateix Bash em dóna
>> errors com «cannot set terminal process group» i que no puc ni
>> instal·lar paquets com ssh.
>> Se m'acudeixen diferents alternatives:
>> - Tornar a Debian 7
>> - Canviar systemd per System-V (si és que es pot)
>> - Que realment hi hagi manera de fer córrer els serveis com abans.
>>
>> Algú em pot aconsellar?
>>
> 

Re: If Linux Is About Choice, Why Then ...

[GiaThnYgeia]

xxx:
> On Tue, Apr 11, 2017 at 6:21 AM, GiaThnYgeia
>> Has Debian always been this crazy and am I so new to this madness?
> 
> Moving the goalposts always generates a bit of madness. Whether this
> time is turning out more so than previous times I'll leave for others
> to comment on.

Moving the goalposts is a decision that weighs on those exact
individuals that made it, not the entire community.

Let me ask this encyclopedic question.  Would newer server hardware
benefit more from the anticipated sysv development or from the systemd?
I suspect that a 4 year old server had a 20% capability improvement with
systemd development while a 1 year old server had a 40% improvement,
despite the fact that the 1 year old one was 4 times as capable as the
older one.  With sysV i suspect the development would maybe affect 24%,
on both.  So was the choice made on the basis of exploiting ever newer
hardware or to continue the human support of those employing Debian for
years?  Disregarding community and bending over to industry, that is!

I have no real data to prove this, I am only suspecting this to be true.

Sometimes sticking to principles and protecting the survivability of the
organization becomes contradictory.  What is meant by "free" is getting
hazier and hazier.  It is now a gray area of market logic.
The question then becomes, as has always been, whose side are you on
boys?  And when I say boys I mean the decision making few that usually
are boys anyway, unlike the woman that wrote the song.


kAt


-- 
 "The most violent element in society is ignorance" rEG

"Who died and made you the superuser?"  Brooklinux

Verizon users verified email list

[elizabeth . smith]

Hi

we would like to learn your interest in acquiring *Verizon* users verified
email list which helps you to improve your business campaign.

*The other telecommunications software users like:* at, sprint, t mobile,
at coverage, Comcast, time warner, sprint coverage, Apple, U.S. Cellular,
Cricket Wireless, Project Fi, Virgin Mobile USA, and many more.

We specialize in providing Decision Maker’s such as C- level, VP level,
Directors and Managers contacts and Technology users list for all
industries across the globe.

* List contains:* Name, Company Name, Phone no, Fax no, Job Title, Email
Address, Complete Mailing address, Revenue Size, Employee Size, Web
Address, etc.

Kindly let us know if you are working on similar requirements.

Anticipate your response

Best Regards,

Elizabeth Smith
Information Consultant

 Reply “remove” to Opt-Out.

Displaymanager resolutie

[Paul van der Vlis]

Hoi allen,

Ik wil de resolutie wijzigen van GDM3, of eventueel een andere
displaymanager zoals LightDM. Maar dit lukt niet zo erg.

Ik heb geprobeerd een regel met xrandr toe te voegen in
/etc/gdm3/Init/Default, zoals in vele manuals staat.
Maar het werkt niet.

Ook zoiets werkt niet:
cp /home/user/.config/monitors.xml /var/lib/gdm3/.config/monitors.xml

Is er hier iemand die hier ervaring mee heeft?

Dit gaat om Debian testing.

Met vriendelijke groet,
Paul van der Vlis.


-- 
Paul van der Vlis Linux systeembeheer Groningen
https://www.vandervlis.nl/

Re: software gezocht

[VinceW]

On 04/11/2017 02:34 PM, lists wrote:
> 
> Bij één of twee projecten is dat wellicht geen groot probleem, maar het
> plan is om er downloads voor verschillende projecten mee te gaan hosten.
> En dan wordt al heel snel wèl problematisch.
> 
> Heeft niemand hier software in gebruik waarmee zoiets kan..?!
> Zo bijzonder lijkt ons deze behoefte toch niet...

Ik ken jullie behoefte qua interface etc niet, maar is git wellicht iets?


Anders zou ik eens kijken naar Drupal. Geen kant en klare applicatie,
maar een solide basis met bruikbare 'bouwblokken'.

Er zijn goed onderhouden contrib modules waarmee je de geschetste
functionaliteit relatief simpel in elkaar kunt steken.

> 
> We willen ook best betalen, het hoeft helemaal niet perse open source /
> gratis te zijn, hoewel dat wel onze voorkeur heeft.

Beiden zijn opensource. Voor beide producten zijn evt. ook nog
commerciele partijen beschikbaar die iets kunnen opzetten/ondersteunen.

> 
> Groet,
> MJ
> 

Best,
VinceW

Re: software gezocht

[Paul van der Vlis]

Op 11-04-17 om 10:16 schreef lists:
> Hallo,
> 
> Wij zijn op zoek naar software om een soort file "download portal" op te
> zetten.
> 
> Ideaal zou zijn wanneer we een webinterface kunnen maken ongeveer als
> 
> http://projecten.company.com/projectX
> http://projecten.company.com/projectY
> 
> Met per project een 'landingpage' met beknopte informatie, en een signup
> button. Als iemand zich bv voor projectX registreert, moet de admin van
> dat project (adminX) een notificatie krijgen, en de aanmelding goedkeuren.
> 
> Vervolgens heeft de aangemelde persoon dus toegang gekregen tot
> http://projecten.company.com/projectX met als doel om daar (vaak grote)
> bestanden te downloaden.
> 
> Optioneel zou ldap / SAML / OpenID Connect voor adminX / adminY logins
> mooi zijn. Ook statistics over downloads is een pre. De zelf-aangemelde
> gebruikers moeten NIET de rechten krijgen om zelf bestanden kunnen
> uploaden en shares aan te maken.
> 
> Producten die we al bekeken hebben zijn: pydio, seafile, own/nextcloud.
> Al deze producten zitten wel in die hoek, maar missen allemaal ook weer
> bepaalde functies.
> 
> Wij denken dat wat we zoeken eigenlijk niet zo bijzonder is,
> desalniettemin kunnen we de perfecte match niet vinden.
> 
> Als iemand tips of suggesties heeft?

Ik heb er weinig ervaring mee, maar kijk eens naar Redmine:
http://www.redmine.org/
https://packages.debian.org/jessie/redmine

Groeten,
Paul

-- 
Paul van der Vlis Linux systeembeheer Groningen
https://www.vandervlis.nl/

Re: If Linux Is About Choice, Why Then ...

[Lisi Reisz]

On Tuesday 11 April 2017 13:54:50 Richard Owlett wrote:
> >> 'They' never told us, owners of single user laptops, why we should chose
> >> it.

Because they don't care whether you chose it or not?  Debian offers 
alternatives, but had to chose a default.  Other distros have made their 
choice.  Ubuntu has chosen - of course - to go where the money is.

No-one cares what Richard Owlett - or Lisi Reisz - has chosen for his (her) 
laptop.  And why should anyone?

Lisi

Re: If Linux Is About Choice, Why Then ...

[Richard Owlett]

On 04/10/2017 07:59 PM, Ric Moore wrote:

On 04/08/2017 01:06 AM, Richard Owlett wrote:


'They' never told us, owners of single user laptops, why we should chose
it.


Simple, as I see it, single user laptop support doesn't pay the bills.
Neither do Desktops. Ubuntu found that one out, for all of their user
friendly features. When Red Hat broke the billion dollar amount for a
single year of support services to big iron, everyone else salivated.
The decision to switch to SystemD is simple ...Docker, clusters and "The
Cloud", where the big dollars roam.



"Simple" it ain't.
I don't believe that the "single user laptop" market has been targeted 
in recent generations of operating systems.


I suspect an underlying problem is "linguistic" rather than "technological".

Ric didn't quote a key (IMMHO;) sentence of my post. I said:

I never saw anything addressing potential advantages for a
specific organic user owning a discrete laptop.


When speaking Linux (as opposed to English) the terms "user" and "single 
user" have meanings and implications quite divergent than what the 
general population would perceive.


The mass marketers [Microsoft, Canonical(Ubuntu), Google(Android)] use 
the "high power blunderbuss" approach. They bag enough sales to keep 
bean counters happy. Please note I did NOT use the term "customer" nor 
"accountants" in the previous sentence.


Now Ric has suggested that Red Hat has developed a large caliber rifle 
known as 'systemd' which is suitable for 'big iron'. The herd of vendors 
have jumped on the band wagon.


To shift idiom a bit. When facing a charging rhino an elephant gun would 
be more useful than a blunderbuss. But when deer hunting, wouldn't a 
hunting rifle with scope be more appropriate than either?


More editorials likely to follow.
The OWL ducks fer cover ;/

Re: Re: how to force mysql to use the name of my machine instead of localhost

[Abdelkader Belahcene]

Thanks for reply;
I give more details.
I install wordpress on my machine, it works fine on local,  I mean i if
write in the browser localhost/wordpress, it 's OK.

I can use myIP/worpress on a remote  machine to connect my server, It seems
working  fine, I get the main page,
BUT, CSS files are not read, I know that by prospecting the code source of
the page.

Wordpress store everything in database,  and so the name of the CSS file
for example, unfortunatly,  the path  used is full address of the file like
localhost/wordpress/blabla/style.css
here is the Big Problem

When the connection is local, no problem, since localhost is the host of
the server;

when the connection is  like :   http://myURL/wordpress,   we SHOULD FIND
the CSS file with  http://myURL/wordpress/blabla/style.css



 To solve the problem

   - either  we remove the absolute path for the files, and keep only  the
   relative part,  instead of storing  localhost/wordpress/blabla/style.css, we
   keep only wordpress/blabla/style.css
   - or we replace the localhost by the name of the machine or IP, that is
   what I want to do, I mean instead of storing in DB, localhost I tore the
   name of my machine like bela.enst.dz, the connetion w'll be done with
   this name and the problem is then solved

I used xampp to install apache, mysql and php. But the problem is the same
when I installed each element alone ( apache2, php, mysql)


when I used the command : localhost -f , I got the correct name
tmp$ hostname -f
bela.enst.dz

thanks a lot for any idea

Regards

bela

Re: Old 32bit PC 650kRam less VidMem 1024x768 will not run on Stretch ok on Jessie

[GiaThnYgeia]

Felix Miata:
> My opinion is that people who want ancient hardware to continue to be
> supported for the longest possible period of time must participate in
> the development phase by testing on the very hardware that they want to
> stay working when it's hardware that developers either no longer have or
> no longer have motivation to use, and they must do so throughout
> development, not only in the final weeks before release is expected.

I'll stick to the "people who want ancient hardware" and ask whether you
perceive those people as having a choice to "want ancient" hardware or
whether this is "all" they have.  Do you anticipate those same people to
be able to start their computing career in developing systems?
Which relates to that world do we want Debian to prevail.  The "free"
world or the "non-free" world in which we live in?

> When a problem is found, it needs to be timely reported according to
> distro policy in the proper place and manner. Reports in the debian-user
> mailing list are insufficient to fulfill this purpose.

By who?  By the person who just lost all graphical access to the system?
By a bug-reporting system that I have yet to see ever working?
Is a person installing debian expected to know how to access and
communicate on the command screen of the -recovery mode?

> You, not we. I'm not here to discuss any policy other than whether this
> thread's content is suitably located, and with this post I'm finished
> with this sub-thread.

Me too ... dealing with such responses

> Did including iomem=relaxed on your cmdline solve your problem, or did
> it not?

I wouldn't know, the person for which I installed Debian for will not
dare switch to Stretch after this experience, will not even talk about
it.  Did you read about the person who used zsh as a login shell and
upgraded Jessie to testing and got locked out of his system because zsh
vanished?  Good thing he knew how to deal with that crisis.

> If yes, say so.

Are you talking to me?  I am one of this privileged well off people to
have hardware that can run Stretch and Sid and all kinds of things.
Send some snail-mail to the less privileged who have lost access to the
net to ask them.

> Supporting ancient hardware forever is not going to happen.

A 2.33Mhz Celeron pc is by no means ancient compared to the Debian claims.

-- 
 "The most violent element in society is ignorance" rEG

"Who died and made you the superuser?"  Brooklinux

Re: software gezocht

[Frans van Berckel]

On Tue, 2017-04-11 at 11:46 +0200, lists wrote:
> On 11-4-2017 10:16, lists wrote:

> > Producten die we al bekeken hebben zijn: pydio, seafile,
> > own/nextcloud.
> > Al deze producten zitten wel in die hoek, maar missen allemaal ook
> > weer bepaalde functies.
> 
> Kijken nu naar ProjectSend. Komt dicht in de buurt van hetgeen we 
> zoeken, maar lijkt een beetje verlaten.

De laatste commit is van 3 april 2017 ;-)

https://github.com/ignacionelson/ProjectSend/commits/master

Met vriendelijke groet,


Frans van Berckel

Fwd: Re: Una de hostapd en debian 7...

[JAP]

(Reenvío a la lista. Colado a mi privado)

On Wed, 29 Mar 2017 18:16:22 -, JAP  
wrote:



El 29/03/17 a las 15:03, luisededios escribió:

Ahora sí caramba!, ya trabaja el AP en el server DELL, y me puedo
conectar desde la cliente Acer, y levanta a mahomacán  :)


¡Bravo-Zulu!


No añado el acostumbrado [SOLUCIONADO] porque ahora quiero hacer los
arreglos para poder compartir la conexión desde el portal cautivo,
además de mejorar todo lo que se pueda en materia de seguridad.

Continúo con mis preguntas mataburro:


Colegas, ahora quiero conectar un nano loco 2 a la laptop Dell, la cual 
ya tiene la app hostapd trabajando.


El nano está configurado en modo estación pero.. cómo puedo, por 
ejemplo, compartir por hostapd(mi AP) lo que recibo, a través, del nano 
en modo estación?  :)


--
Saludos,
Luis

Re: Old 32bit PC 650kRam less VidMem 1024x768 will not run on Stretch ok on Jessie

[GiaThnYgeia]

deloptes:
> songbird wrote:
> 
>> did you check /var/log/Xorg.0.log?
> 
> I doubt he does know about it

You doubt or you know?
I knew about it but I did not check it.  There are many logs on /var/log
and its subdirectories.  Can I interpret 3% of what they say?  No!  Do I
need to in order to install and run a system instead of mswin?  I hope not!

Once the final re-installation was done with Jessie what purpose would
checking the log in a functional system?  The log of the system that
would not result in a functional graphic display would probably show
something meaningful.

I did not think this way, I just thought the display itself being old
and of low resolution had come to an altered default resolution that it
could not handle.  I have seen this on other debian based distros, where
the live or installer resolution was so off that you couldn't see half
the installation window.  The debian installers have generally faired
well in this respect.  Although I have had successful installations with
TV screens and parts of the Debian installer off the frame.

Why would a functional system in Jessie be allowed to be broken by a
simple upgrade without at least a warning.  "Your hardware is not
supported by testing/stretch.  If you continue with the upgrade you may
lose graphic functionality of the system".  Shouldn't the same people
who made the decision not to continue the support be responsible to make
sure a patch exists in Jessie that adds such a warning to the user?
Once deleting and reinstalling various display managers and failing, I
scrapped it and started with a new Jessie.

Your excuse for "everything is just fine" holds no water at all for
anyone who does not know of Xorg.0.log unless you really think that
anyone has memorized the Debian administrator's manual in its entirety
BEFORE they attempt an installation.

But this is a users discussion list, not developers.  All we need is a
functional system to do web-browsing, word-processing, database for home
or store use, and a card game.  USERS are not those who are expected to
diagnose and repair developers' broken systems!

And please refrain from giving me a "testing" not "stable" speech as it
is not May 2015, it is almost 2 years later under a long standing
freeze.  And don't dare imply that in stable stretch issues like this
will be taken care of without substantial evidence for this claim.

https://bugs.debian.org/release-critical/graph-release.png
https://bugs.debian.org/release-critical/

Look at this misleading chart, Jessie is blue, Stretch is green, which
would you pick to start your experience with Linux?  The neglected
Jessie or the hot off the press Stretch?  Stretch should really be
Jessie 8.8  but 8.8 should not break your display management.

kAtrinity

-- 
 "The most violent element in society is ignorance" rEG

"Who died and made you the superuser?"  Brooklinux

Re: software gezocht

[lists]

On 11-4-2017 10:16, lists wrote:

Producten die we al bekeken hebben zijn: pydio, seafile, own/nextcloud.
Al deze producten zitten wel in die hoek, maar missen allemaal ook weer
bepaalde functies.


Kijken nu naar ProjectSend. Komt dicht in de buurt van hetgeen we 
zoeken, maar lijkt een beetje verlaten.


MJ

Re: (deb-cat) Chroot des d'Android

[Narcis Garcia]

He provat a canviar systemd per SystemV al contenidor:
$ apt-get install sysvinit-core sysvinit-utils
$ cp /usr/share/sysvinit/inittab /etc/
$ apt-get remove --purge --auto-remove systemd

Però segueixo tinc problemes quan intento instal·lar altres paquets:
dpkg (subprocess): cannot set security execution context for maintainer
script: Invalid argument

he llegit que això té a veure amb SELinux, i n'he instal·lat les
utilitats per provar de deshabilitar-lo:
$ apt-get install selinux-basics auditd selinux-utils
$ setenforce 0
no sembla tenir cap efecte potser perquè el nucli Linux d'Android ja ha
arrencat amb això;
$ selinuxenabled && echo enabled || echo disabled
enabled

NOTA: Ja he vist les aplicacions «Debian Kit» i «Lil' Debi» de
f-droid.org, però el problema és que treballen amb una imatge de disc
virtual, i per aquest fitxer cal menjar-se bastant espai
d'emmagatzematge. Jo prefereixo compartir l'espai lliure amb la resta
d'entorn.


__
I'm using this express-made address because personal addresses aren't
masked enough at this list's archives. Mailing lists service
administrator should fix this.
El 10/04/17 a les 14:16, Narcis Garcia ha escrit:
> Estic preparant una tauleta per ensenyar una aplicació web sense
> necessitat d'accés a internet.
> 
> Per això he alliberat els permisos d'Android (root) i hi he instal·lat
> una Debian 8 dins un directori amb debootstrap.
> Per iniciar-hi serveis en un entorn així, sempre n'havia tingut prou amb
> executar-los cuidant una mica les variables d'entorn i punts de muntatge
> (fins i tot servia per iniciar escriptoris).
> 
> No sé si és gràcies al venerable systemd que el mateix Bash em dóna
> errors com «cannot set terminal process group» i que no puc ni
> instal·lar paquets com ssh.
> Se m'acudeixen diferents alternatives:
> - Tornar a Debian 7
> - Canviar systemd per System-V (si és que es pot)
> - Que realment hi hagi manera de fer córrer els serveis com abans.
> 
> Algú em pot aconsellar?
> 

software gezocht

[lists]

Hallo,

Wij zijn op zoek naar software om een soort file "download portal" op te 
zetten.


Ideaal zou zijn wanneer we een webinterface kunnen maken ongeveer als

http://projecten.company.com/projectX
http://projecten.company.com/projectY

Met per project een 'landingpage' met beknopte informatie, en een signup 
button. Als iemand zich bv voor projectX registreert, moet de admin van 
dat project (adminX) een notificatie krijgen, en de aanmelding goedkeuren.


Vervolgens heeft de aangemelde persoon dus toegang gekregen tot 
http://projecten.company.com/projectX met als doel om daar (vaak grote) 
bestanden te downloaden.


Optioneel zou ldap / SAML / OpenID Connect voor adminX / adminY logins 
mooi zijn. Ook statistics over downloads is een pre. De zelf-aangemelde 
gebruikers moeten NIET de rechten krijgen om zelf bestanden kunnen 
uploaden en shares aan te maken.


Producten die we al bekeken hebben zijn: pydio, seafile, own/nextcloud. 
Al deze producten zitten wel in die hoek, maar missen allemaal ook weer 
bepaalde functies.


Wij denken dat wat we zoeken eigenlijk niet zo bijzonder is, 
desalniettemin kunnen we de perfecte match niet vinden.


Als iemand tips of suggesties heeft?

Bedankt!

MJ