Re: rsync to NAS for backup

[Mirko Parthey]

On Mon, Feb 15, 2021 at 12:39:29PM +, mick crane wrote:
> Appears that to retain permissions need root at both ends of rsync.

Not necessarily. If the server filesystem supports xattrs, you can use
the --fake-super option with the rsync server, running as a non-root user
that can write to the storage.

To set the option on the client side for rsync over ssh, use:

--rsync-path "rsync --fake-super"

Regards,
Mirko

Re: OT: Router behaviour

[Mirko Parthey]

On Thu, Feb 04, 2021 at 11:03:17PM +, Mark Fletcher wrote:
> My kids have been complaining recently about the quality of the WiFi

I suggest to investigate whether the problem lies with WiFi or elsewhere.

There are multiple tools available for measuring throughput and latency.
I use flent with netperf, and for the Internet connection also
http://www.dslreports.com/speedtest

IMHO the best way to improve a weak WiFi signal is to install multiple
access points throughout the house and connect them by ethernet cable to
the main router.

Some latency problems can also be fixed with alternative firmware and
configuration on the router.

> I powered the router down again, plugged its WAN port into one of the
> LAN ports of the ISP-supplied router, and brought it back up. It came up
> but seems to have automatically subordinated itself to the ISP-supplied
> router and is now offering up IP addresses in the range supplied by the
> ISP's router...

The secondary router could be in bridge mode during startup,
passing DHCP from its LAN clients to the main router and back.

See for example the following discussion in the OpenWrt forum:
https://forum.openwrt.org/t/bug-in-the-dhcp-mechanism/14777

Often this problem cannot be truly fixed without buying new hardware,
but there are workarounds if you do not mind reconfiguring the network.

For example, you could set up your secondary router as an access point,
if its firmware allows it:

- disable its DHCP server
- configure its LAN to use a static IP address compatible with the main
  router's subnet, not used by another device yet and outside the main
  router's DHCP range
- alternatively, configure its LAN as a DHCP client (if possible at all)
- connect it to the main router using its LAN port
  (its WAN remains disconnected)

Regards,
Mirko

Re: Encountered a bug with a dependency of wondershaper, but I'm unsure which dependency, and how to proceed with submitting a bug report

[Mirko Parthey]

On Tue, Nov 17, 2020 at 09:15:15PM +0100, Graham Bull wrote:
> I've noticed when I set the same rules within wondershaper on Stable and
> Testing, I get different behavior.
> Stable acts as expected, low latency and able to hit the limits set.
> Testing suffers a lot of latency and I'm only able to reach a fraction of the
> limit set.
>
> When I remove the wondershaper rules everything works as expected (I can max 
> my
> internet connection on both computers).
>
> [...]
>
> Any advice of how to collect more info for debugging purposes or how to 
> proceed
> would be very much appreciated!

Short summary: I suggest to dump wondershaper and use CAKE instead.
Please read on for details.

On typical ISP access links, there are devices on both ends of the link
that introduce high network latency under load
(xDSL modem, DSLAM, cable modem, CMTS).

This can be shown well by the DSLreports speed test:
http://www.dslreports.com/speedtest

Often, these devices use large network buffers and too simple buffer
management, which cannot be changed by the user.
Your only chance to reduce the latency is to set up a traffic shaper
on a device under your control to make it a "choke point".
I guess you know this since you are using wondershaper.

A good solution is the combination of a shaper (e.g. HTB) with fq_codel.
wondershaper does not use fq_codel yet.

Personally, I use the CAKE qdisc since it is easier to set up and comes
with even more improvements over HTB + fq_codel:
  man tc-cake
  https://www.bufferbloat.net/projects/codel/wiki/Cake/

My router runs the OpenWrt distribution, but CAKE should also work on Debian.

Before giving further advice, I would like to know more about
your network:
- Does the Debian box in question manage your Internet access link, or
- is there a separate router, and are there other devices using it?
- What is the type (xDSL, cable, ...) and speed of your Internet link (down/up)?

Regards,
Mirko

Re: slow wireless connection

[Mirko Parthey]

On Sun, Aug 11, 2019 at 01:11:43PM -0400, kamaraju kusumanchi wrote:
> The internet download speed as measured by (speedtest.net) is ~15 Mbps
> when I try to connect from my desktop. From a different machine (my
> laptop), I get around ~30-40 Mbps. Could you please tell me how to fix
> this?
>
> Network connection:
> {desktop or laptop) connected wireless to -> linksys wrt54g router ->
> ISP (optimum)

Which model of the WRT54G do you have?
https://wikidevi.com/wiki/Linksys_WRT54G_series

Is it running the vendor firmware or some other such as OpenWrt?

All models of the WRT54G are slow by today's standards, and are likely
to be a bottleneck for network throughput, even at such moderate bitrates.
This doesn't explain the difference between the two clients, though.

> On the desktop, I am using "Panda Ultra 150Mbps Wireless N USB
> Adapter"  to connect to
> the router.
>
> The laptop has a builtin wifi adapter, Intel(R) Dual Band Wireless-AC 8265.
>
> The desktop is running Debian stretch. The laptop is running Windows 10 Home.

I'd suggest plugging the USB adapter into the laptop and comparing the
throughput of the built-in adapter vs. the USB one, by enabling each of
the wireless interfaces in turn and disabling the other.
Repeat the test with a linux live system, also on the laptop.
This should reveal whether the USB wireless adapter is the bottleneck.

Which frequency band / channel are you using?
The USB plug seems to be physically small, without enough space to fit a
reasonably-sized 2.4 GHz antenna. It might work better on 5 GHz,
but I am not aware of any linksys wrt54g model which supports this band.

Regards
Mirko

Re: Apache HTP-Server won't fork more than 3000 processes.

[Mirko Parthey]

On Tue, Sep 18, 2018 at 12:37:48PM +0200, Martin wrote:
> > # cd /sys/fs/cgroup/pids/...
> > # cat pids.max
> 
> There is no pids.max
> But I have a /proc/sys/kernel/pid_max = 32768
> I guess like systemd is limiting here?
> 
> > How many tasks are assigned to the cgroup?
> > # wc -l tasks
> 
> 103 tasks

On a Debian stretch VM (32 bit), I found:

root@stretch32:/sys/fs/cgroup/pids/system.slice/apache2.service# cat pids.max 
4915

root@stretch32:/sys/fs/cgroup/pids/system.slice/apache2.service# wc -l tasks 
55 tasks

Re: Apache HTP-Server won't fork more than 3000 processes.

[Mirko Parthey]

On Tue, Sep 18, 2018 at 12:14:11AM +0200, Martin wrote:
> I have an issue with Apache's HTTP-Server 2.4. It says 'Resource temporarily 
> unavailable: AH00159: fork: Unable to fork new process'. I know, there are 
> some hits with Google, but there is no solution to me.
> Systemd knows about a task limit of 6000. mpm_prefork_module is set to 
> 'ServerLimit 4000' and 'MaxRequestWorkers 4000'. ulimits are set to unlimited.
> I also see in the kernel log 'cgroup: fork rejected by pids controller in 
> /system.slice/apache2.service'. But this does not correlate with the httpd's 
> messages from above, that comes only from time to time.
> Why does this thing stop forking at about 3000 processes? I have a (not just 
> one) Red Hat 6 in parallel, that one does.

There may be two issues here, my answer refers to 
'cgroup: fork rejected by pids controller'.

Is there any limit set in pids.max for the apache cgroup under the pids 
controller?
# cd /sys/fs/cgroup/pids/...
# cat pids.max

How many tasks are assigned to the cgroup?
# wc -l tasks

Mirko

Re: Local system program analytics

[Mirko Parthey]

On Thu, Jul 12, 2018 at 10:09:10AM +0200, to...@tuxteam.de wrote:
> So the programs which burn most CPU will get most sponsoring ;-)

Maybe wall-clock time spent is a suitable measure?
sa --sort-real-time | head

> I just wanted to illustrate how difficult a "good" answer to the
> OP's seemingly simple question is...

I agree.

Mirko

Re: Local system program analytics

[Mirko Parthey]

On Wed, Jul 11, 2018 at 07:18:55AM -0400, David Maulik wrote:
> Is there any program that can show me the programs on my system I use the 
> most?
> I'm looking for something that can help me trim unneeded/unused packages as
> well as show me which programs I use the most to help inform support/donation
> decisions.

You could try BSD process accounting. The Debian kernels have it built in,
and the userspace tools such as "sa" can be found in the "acct" package.

Re: new install of amd64, 9-4 from iso #1

[Mirko Parthey]

On Sun, Jun 10, 2018 at 04:44:16PM -0400, Gene Heskett wrote:
> and 3: to  treat the grub install as if there are no other drives hooked 
> up. I don't need grub to fill half the boot screen with data from the 
> other drives.

Once your Debian installation is finished, put this in /etc/default/grub:
GRUB_DISABLE_OS_PROBER=true

Then run
  update-grub
to remove the unwanted entries from your grub menu.

The Grub info documentation describes it as follows:
'GRUB_DISABLE_OS_PROBER'
 Normally, 'grub-mkconfig' will try to use the external 'os-prober'
 program, if installed, to discover other operating systems
 installed on the same system and generate appropriate menu entries
 for them.  Set this option to 'true' to disable this.

Disconnecting disks while installing Debian can help avoid mistakes.
However, it does not permanently suppress the boot menu entries referring to
other OS installations.

Regards,
Mirko

Re: Any Sound Card Recommendations?

[Mirko Parthey]

On Tue, Nov 21, 2017 at 09:42:01AM -0500, Thomas George wrote:
> I bought an Asus Xonar DSX expecting it to work with my Debian Stretch
> system, there was a recommendation on the Internet from someone using Mint
> who said it worked right out of the box. Mine didn't.
> 
> I would like a really good sound card for my system and don't want to make
> the same mistake twice.
> 
> Any recommendations?

What does "really good" mean to you?
Please tell us more about your requirements, such as:

* usage scenario (music listening, recording, games, home theatre, ...)
* PC-side interface (PCIe, USB)
* number of audio channels in/out needed and their signal type
  (microphone, phono, line, headphones, digital interfaces)
* sample frequency / bit depth
* signal processing on the soundcard (MIDI to PCM, Dolby whatever, ...)
  if needed at all

Re: GPG-Error with debmirror and wheezy-proposed-updates

[Mirko Parthey]

On Fri, Jul 28, 2017 at 05:14:20PM +0200, Christoph Pleger wrote:
> [GNUPG:] NEWSIG
> [GNUPG:] KEY_CONSIDERED A1BD8E9D78F7FE5C3E65D8AF8B48AD6246925553 0
> [GNUPG:] KEY_CONSIDERED A1BD8E9D78F7FE5C3E65D8AF8B48AD6246925553 0
> [GNUPG:] BADSIG 8B48AD6246925553 Debian Archive Automatic Signing Key 
> (7.0/wheezy) 
> gpgv: Signature made Do 20 Jul 2017 16:31:09 CEST
> gpgv:using RSA key 8B48AD6246925553
> gpgv: BAD signature from "Debian Archive Automatic Signing Key (7.0/wheezy) 
> 
> .temp/.tmp/dists/wheezy-proposed-updates/Release.gpg signature does not 
> verify.
> 
> Does anyone have an idea what is the cause of this? Is it on the client
> side, that is, on my mirror computer? Or on the server side, that is, on
> ftp.us.debian.org? And what do I have to do to solve this on the client or
> let it solve on the server?

Looks like there is a broken GPG signature on the Debian FTP server.
To contact the ftpmaster team:
https://ftp-master.debian.org/index.html#contact

Mirko

Re: debmirror, GPG, Stretch

[Mirko Parthey]

On Thu, May 11, 2017 at 03:21:17PM +0200, Andre Majorel wrote:
> This is where a list of key IDs used for ftp.debian.org would
> come in handy. But if it exists, it's not easy to find.
> 
> No doubt keys can be found by looking into the latest release of
> debian-archive-keyring or poking strings into a search engine
> until you get lucky. What I'd like to know is where are we
> *supposed* to look for them ?

Individual archive keys are available here:
https://ftp-master.debian.org/keys.html

Mirko

Re: Failing disk advice

[Mirko Parthey]

On Sun, Mar 05, 2017 at 08:38:27PM -0800, David Christensen wrote:
> On 03/05/2017 01:02 PM, Gregory Seidman wrote:
> >I have a disk that is reporting SMART errors. It is an active disk in a
> >(kernel, not hardware) RAID1 configuration. I also have a hot spare in the
> >RAID1, and md hasn't decided it should fail the disk and switch to the hot
> >spare. Should I proactively tell md to fail the disk (and let the hot spare
> >take over), or should I just wait until md notices a problem?
> 
> I'm confused by "I also have a hot spare in the RAID1".  Do you have a
> two-member RAID1 with a hot spare, or a three-member RAID1?  I would prefer
> the latter:
> 
> https://manpages.debian.org/jessie/mdadm/md.4.en.html

Refining this advice a bit, I would convert the spare to a full RAID
member now, without explicitly failing the disk that reports SMART
errors first.
Assuming you have a two-member RAID1 with a hot spare, the command
should be similar to this (untested):
  mdadm -G /dev/mdX -n 3 
This ensures you keep redundancy during further maintenance actions.

Which SMART errors do you get, and who reports them?
What is the output of the following command for the failing drive?
  smartctl -A /dev/sdY

Regards,
Mirko

Re: Dell BIOS

[Mirko Parthey]

On Mon, Jan 16, 2017 at 11:52:48AM -0600, Martin McCormick wrote:
> change the boot sequence order from floppy-C:-CDROM to
> CDROM-floppy-C: or CDROM-C:-floppy which prevents the hard drive
> from grabbing the boot sequence each time.

There may be a function key that brings up a BIOS boot menu.
For Dell machines, it should be F12. Other candidates are F8 and F11.
Choosing an entry in the menu takes effect exactly once,
then the previous boot order is restored.
This feature might be useful to you if the entries are the same each
time and appear in the same order.

>   I don't think the CMOS batteries are dead because the
> clock still seems to keep time

Changing the battery is certainly worth trying.

The BIOS clock could still have the correct time despite a dead battery
because the Linux system gets the time from an NTP server and writes it
to the BIOS clock on shutdown.  The clock then runs on ATX standby power
and will only lose its memory when the power cable is removed. The same
applies to the BIOS settings.

Regards,
Mirko

Re: Low Level Format of 1.4 Mb Floppy Isn't Happening

[Mirko Parthey]

On Sat, Jan 14, 2017 at 03:32:08PM -0600, Martin McCormick wrote:
> What happened after I zapped the floppy is that fdformat will not
> run because it sees no pre-existing format information.
> 
>   What am I forgetting or what has changed?

According to the fdformat man page, you should first set the disk
parameters with setfdprm, or use a device like /dev/fd0H1440 which
refers to a specific format.

If any of the device nodes are missing, you can create them with mknod,
the major and minor numbers are given in the fdformat man page as well.

Other options to boot an old machine could be
a Compact Flash card with a CF-to-IDE adapter,
or an Ethernet card with PXE support and a BOOTP/DHCP and TFTP server.

Regards,
Mirko

Re: debian on lenovo carbon thinkpad 4th generation confirmation report

[Mirko Parthey]

On Tue, Nov 29, 2016 at 10:59:35PM +, Lisi Reisz wrote:
> > I didn't know that kernels > 4.1 are available as jessie-backports. Thanks.
> > Regards, Karol
> 
> Only 4.7.  The others appear not to be available any more.

Older backports are available at http://snapshot.debian.org/.

Mirko

Re: Does hdparm not run at startup anymore?

[Mirko Parthey]

On Fri, Nov 11, 2016 at 11:11:24PM +0100, Rainer Dorsch wrote:
> I configure sdb in /etc/hdparm.conf to apm=64, but when I start the system, 
> apm
> does not change. Interesting enough a /etc/init.d/hdparm restart fixes the
> problem:

There are two config options available: "apm" and "apm_battery".
Which one takes effect depends on your machine's power status:
connected to external power or running on battery.
Does this help you?
If not, please run these commands as root and show us their output:

  find /etc/rc* \( -name "*hdparm" -o -lname "*hdparm" \) -ls
  ls -l /sbin/init
  systemctl status hdparm

Regards
Mirko

Re: Reuniting disks in a raid1 array

[Mirko Parthey]

On Tue, Aug 30, 2016 at 07:33:02AM +0200, Frédéric Marchal wrote:
> I had two disks in a mdadm software raid1 on an old computer.
> 
> I moved sda to a new computer. It ran there in a degraded raid1 for months.
> 
> The second disk, sdb, kept running as the lone survivor of the original raid1 
> on the old computer.
> 
> Now, it's time for the old computer to retire and sdb to join its partner in 
> the new raid1 on the new computer.
> 
> How can I do this safely?
> 
> Do I have to wipe sdb before adding it to the new computer? If so, how do I 
> make sure raid data is gone from every one of the three partitions?

The wipefs tool can remove RAID signatures.

Deleting and recreating the partitions with the same size doesn't help
on its own, because it will not affect the signatures.
Overwriting the beginning of the disk or the partitions may not be
sufficient either, because depending on the RAID superblock version, it
might be stored at the end.

> Or maybe mdadm will see that sda was used more recently than sdb and will 
> synchronize sda onto sdb?

I would advise against such experiments. It's best to wipe the signatures
while the disk is still in the old computer, then transfer it to the new
one. A bootable rescue system such as GRML can be very handy for this job.

Regards
Mirko

Re: using debian debootstrap and chroot to login to other distro on neighbouring partition etc

[Mirko Parthey]

On Thu, Feb 04, 2016 at 06:05:50PM -0600, Dutch Ingraham wrote:
> I'm not sure debootstrap is what you are looking for here.  If you just 
> want to chroot into your Ubuntu,on the same disk, these are the steps:
> 
> 1. Make a mount point, say /mnt/ubuntu;
> 
> 2. Mount the partition Ubuntu is on, e.g., ;
> 
> 3. Change directory to /mnt/ubuntu;
> 
> 4. ;
> 
> 5. ;
> 
> 6. ;
> 
> 7. ;
> 
> 8. ;
> 
> 9. ;
> 
> 10. ;
> 
> 11. ;
> 
> 12.  
> Of course, you will need to determine certain things up front and modify for
> your particular needs, i.e., which partition Ubuntu is currently residing on,
> whether you need network access, etc. Some of these commands will need to
> be customized/omitted based upon those needs.

This is good advice, and these steps can be automated with the schroot package.

For any non-trivial operations inside your guest systems, such as
installing packages or running daemons, I can recommend Linux containers
(lxc or libvirt-lxc). They look like a fully booted Linux system, but
have low virtualization overhead because they use the host's kernel,
still they have better isolation from the host system compared to the
chroot solution.

Regards,
Mirko

Re: using debian debootstrap and chroot to login to other distro on neighbouring partition etc

[Mirko Parthey]

On Fri, Feb 05, 2016 at 12:49:59PM +, Michael Fothergill wrote:
> This is good advice, and these steps can be automated with the schroot
> package.
> 
> For any non-trivial operations inside your guest systems, such as
> installing packages or running daemons, I can recommend Linux containers
> (lxc or libvirt-lxc).
> 
> 
> I want to install openrc in an install where it can't boot up because I
> deleted it by mistake.  If I use e.g. lxc how many of the above commands do I
> need to enter into it - does it sniff out existing partitions and create
> appropriate mount points by itself etc?

lxc doesn't help if the boot process of your guest is broken.
I would recommend chroot for the repair work.
For a one-off job, schroot might not even be worth setting up.

If you want to try an lxc variant once your guest sytem is fixed:
Most of the mount commands from the chroot solution are not needed with
lxc because it runs the guest system's boot process, which mounts the
necessary filesystems based on the guest's /etc/fstab and other knowledge
hard-coded inside the guest system.

Guest machines for libvirt-lxc can be easily set up using virt-manager,
that's what I used here.
It is my understanding that plain lxc offers more features and has to be
set up manually with a config file, but I haven't used it yet.

Mirko

Re: Do I need to do fstrim on Jessie?

[Mirko Parthey]

On Wed, Nov 25, 2015 at 10:29:26AM +0200, David Baron wrote:
> > > Either you add "discard" as a mount option to your fstab or you crate a
> > > cronjob to run "fstrim -a -v".
> > 
> 
> fstab created by the installer (normal SATA HDs) only has "defaults." Is this 
> implicit here?

You don't need "discard" for normal SATA HDDs, since they don't have a
wear-levelling layer which would manage free space below the filesystem.

>From the mount(8) manpage:

  The discard function issues frequent commands to let the block device
  reclaim space freed by the filesystem. This is useful for SSD devices,
  thinly provisioned LUNs and virtual machine images, but may have a sig‐
  nificant performance impact.

Regards,
Mirko

Re: missing directory /lib/modules/3.16.0-4-amd64/build

[Mirko Parthey]

On Mon, Nov 02, 2015 at 12:18:33AM +0600, EVGeny Dedov wrote:
> After reinstall debian I tried to put wireless drivers in OS by usual way 
> (like
> i did it before), and faced with problem when I use "make compile" and "make
> install":
> 
> make[2]: *** /lib/modules/3.16.0-4-amd64/build: Нет такого файла или каталога.
> (error message translation: "no such file or directory")
> [...]
> what's the problem?
> Thanks for attention.

Install the package linux-headers-3.16.0-4-amd64.

Regards
Mirko

Re: impending disk failure?

[Mirko Parthey]

On Sat, Oct 17, 2015 at 02:15:52PM +0100, Tony van der Hoff wrote:
> Hi,
> 
> T'm occasionally getting this message in syslog on my jessie box:
> 
> Oct 17 12:00:19 tony-lx kernel: [ 8838.600489] ata3.00: exception Emask 0x10
> SAct 0x10 SErr 0x40 action 0x6 frozen
> Oct 17 12:00:19 tony-lx kernel: [ 8838.600501] ata3.00: irq_stat 0x0800,
> interface fatal error
> Oct 17 12:00:19 tony-lx kernel: [ 8838.600509] ata3: SError: { Handshk }
> Oct 17 12:00:19 tony-lx kernel: [ 8838.600517] ata3.00: failed command:
> WRITE FPDMA QUEUED
> Oct 17 12:00:19 tony-lx kernel: [ 8838.600531] ata3.00: cmd
> 61/c0:20:90:6e:eb/01:00:22:00:00/40 tag 4 ncq 229376 out
> Oct 17 12:00:19 tony-lx kernel: [ 8838.600531]  res
> 40/00:20:90:6e:eb/00:00:22:00:00/40 Emask 0x10 (ATA bus error)
> Oct 17 12:00:19 tony-lx kernel: [ 8838.600538] ata3.00: status: { DRDY }
> Oct 17 12:00:19 tony-lx kernel: [ 8838.600547] ata3: hard resetting link
> Oct 17 12:00:20 tony-lx kernel: [ 8839.092521] ata3: SATA link up 6.0 Gbps
> (SStatus 133 SControl 300)
> Oct 17 12:00:20 tony-lx kernel: [ 8839.099092] ata3.00: configured for
> UDMA/133
> Oct 17 12:00:20 tony-lx kernel: [ 8839.099121] ata3: EH complete
> 
> 
> Can anyone please explain what it means, and whether I should be worried?

To find out if this error is caused by the harddisk or by other components,
I'd suggest moving the disk to a different computer.

If the errors remain, your disk is failing.
Otherwise, swap components one by one.
A good next candidate would be the power supply unit.
Also, check the mainboard capacitors for a bulged top or leakage.

Regards,
Mirko

Re: Deleting i386 packages

[Mirko Parthey]

On Mon, Sep 21, 2015 at 09:18:01PM -0600, Joe Pfeiffer wrote:
> For historical reasons, my x86-64 architecture computers have a large
> number of i386 packages on them that I'd just as soon be rid of.  is
> there a good way to simply tell a package manager that I want everything
> involving that architecture deleted?  The best answer I've found on my
> own has been to use dpkg and grep to find everything with :i386, and
> then construct a huge dpkg --purge command to get rid of them all.
> Hoping for something a little simpler...

# aptitude --purge-unused markauto "~ri386"

This should only remove packages that are not needed anymore,
assuming that nobody messed up the auto markings of your installed
packages.

At the yes/no prompt, you might want to check the list of packages to be
purged.
You could also run a first pass without "--purge-unused", check if you are
happy with the results, then do a second pass with this option included.

Regards,
Mirko

Re: PAE

[Mirko Parthey]

On Mon, Aug 24, 2015 at 08:44:44PM -0400, Stephen Powell wrote:
 PAE may be a necessary condition for NX, but it is not a sufficient condition.
 I am presently using three 32-bit computers: one is a Pentium M (2G), one is
 a Pentium 4 (2G), and one is a Xeon (4G).  All three are PAE-capable, and all
 three are presently running PAE Linux kernels.  And all three display the
 following message during boot, according to dmesg|less:
 
Notice: NX (Execute Disable) protection missing in CPU!
 
 So, for my hardware, that argument doesn't seem to hold up.  Am I missing
 something?

No, you are right. I even run such a machine myself!

Regards,
Mirko

PAE (was: Debian Gnome Or XFCE ?)

[Mirko Parthey]

On Mon, Aug 24, 2015 at 06:45:10AM -0400, Stephen Powell wrote:
 Furtherfore, a non-PAE kernel is
 useful even on PAE-capable hardware.  The main purpose of PAE is to
 address memory above 4G.  But if the machine has less than 4G of
 memory, what does a PAE-capable kernel buy you?  PAE-capable kernels
 tend to be a bit bigger, all other things being equal, than non-PAE
 kernels, which chews up more precious memory with no obvious benefit.

One such benefit is that the NX bit (non-executable memory pages)
is only available with 64 bit page table entries, which in turn depend on
PAE mode. This could be an argument for preferring a PAE kernel on
PAE-capable hardware.

Regards,
Mirko

Re: creating a simple private repository with reprepro

[Mirko Parthey]

On Sun, Mar 01, 2015 at 11:03:30AM +0200, Asaf Dalet wrote:
 I have a private debian package which I want publish in binary format.
 I want to supply 2 binary versions:
 1. precise/amd64
 2. trusty/amd64
 
 [...]
 
 I read somewhere that reprepro does not support having 2 files with the same
 name in a repo. is it true?

Yes.

 So, my questiona are eventually:
 1. is there a way for me to use reprepro without changing my .deb file names?
 2. if not, what's the simplest alternative?

I would recommend to give your packages different revision numbers
in debian/changelog before building them.
For example, append '+precise' or '+trusty' to the string inside '()'.
This results in different names for the .deb files and is sufficient
to prevent the name collision in reprepro's pool directory.

Alternatively, only build your package for the older distribution and
then try to use it on the newer one. If it works, you can include the
same package file in both distributions, and reprepro should stop
complaining.

Finally, you could set up two independent reprepro repositories,
but I would only do this if the distributions are also maintained
independently (lets say Debian and Ubuntu).

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20150301130647.GA27987@guitar.localdomain

Re: creating a simple private repository with reprepro

[Mirko Parthey]

On Sun, Mar 01, 2015 at 03:18:16PM +0200, Asaf Dalet wrote:
 Mirko,
 As for your first suggestion - do you imply that I change the changelog before
 each run of debuild/pbuilder?

Yes, at least every time the upstream version changes, or every time you
build for a different distribution. You can use debchange -l to automate
it somewhat.

Regards,
Mirko

PS: Sorry for answering off-list last time.
Re-adding Cc: debian-user@lists.debian.org.


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20150301133526.GB27987@guitar.localdomain

Re: disk (and other resource) management of virtual machines

[Mirko Parthey]

On Fri, Dec 19, 2014 at 02:19:18PM -0800, Ross Boylan wrote:
 Is there a way to expose host file systems to the guests?  NFS is a
 possibility, but the VM's will be running various services that warn
 not to use NFS.  libvirt doesn't seem to provide the ability to expose
 host file systems directly (as opposed to exposing raw block devices),
 but I'm hoping I have missed something.

http://www.linux-kvm.org/page/9p_virtio
I have no experience with it, so please evaluate it yourself.

 I could run LVM inside the VM's; this would provide somewhat more
 flexibility and disk snapshotting, though I'm not sure how I would
 provide more total space to the VM in this scenario.

To extend the space available to the guest in this scenario,
you need to work layer by layer.
On the host:
- enlarge the logical volume containing the guest image
In the guest sytem:
- enlarge or add a partition containing a physical volume
- pvresize or pvcreate/vgextend to enlarge the volume group
- lvresize -r to enlarge the logical volume and the filesystem;
  or fsadm (same purpose)

 Finally, could anyone clarify the VMs in this libvirt/KVM setup use
 CPU and RAM resources?  If I give a guest machine 4G of RAM does that
 mean that memory is unavailable for other use by the host or other
 guests? virt-manager has configuration for memory and maximum memory,
 suggesting that at least some dynamic growth is possible.  Likewise,
 if the physical machine has 8 cores, could I run several VM's with 8
 cores each?  My suspicion is that CPU's are shareable, but RAM is not,
 but I don't know.

Yes, CPUs are shareable.
How many CPUs should be assigned to each guest depends on the number of
guests and their CPU usage. I would assign all cores to a single
guest only if the load is very light. With higher CPU load, guests
should be assigned enough CPUs to keep all physical CPU cores busy,
and maybe a little more to allow for scheduling flexibility.
Above a certain number of CPUs assigned, you will just increase the
scheduling overhead, but not gain better performance.

RAM is shareable in theory, but current operating systems use up all
available RAM as disk buffers, so some tuning may be necessary.
One possibility is the current memory / maximum memory setting you
mentioned, which needs a guest driver for ballooning.
The guest driver reserves some memory from the guest OS for exclusive
use (the difference max - current).  It doesn't actually use that
memory, but returns it to the host so it can be used elsewhere.
I don't know if this RAM reassignment has been automated yet to change
based on demand.

 Thanks.
 Ross Boylan
 
 P.S. I don't need to be able to make the changes while the VM's are
 live, though the ability to do so would be handy.

If you can take your VMs offline, there is another option for accessing
the guest filesystems: http://libguestfs.org/

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141220141352.GA4750@guitar.localdomain

Re: Using PuTTY with Debian GNU/Linux Systems

[Mirko Parthey]

On Sat, Oct 11, 2014 at 07:34:06AM -0400, Stephen Powell wrote:
 
 In response to encouragement from several people on this list,
 I have published a new web page titled Using PuTTY with Debian
 GNU/Linux Systems.  It is available here:
 
http://users.wowway.com/~zlinuxman/putty.htm
 
 All feedback, both positive and negative, is welcome.  In particular,
 if there are any factual errors on this page, I especially want to
 know about that.

Hello Stephen,

thanks for your web page. I like how you research things in detail,
also enjoyed your work on custom Linux kernels and their interaction
with bootloader configuration some years ago.

In your current article, you suggest rebuilding ncurses to add a
custom terminal type. I find it much easier to just install a terminal
definition locally as a configuration file.

This can be achieved with the following command:
  tic -x terminaldef.src
where terminaldef.src contains just the single entry from your patch.

The output is written to /etc/terminfo/ or $HOME/.terminfo/,
depending on the privileges of the user the tic command runs as.
The ncurses library searches for terminal definitions in several places,
including these two directories.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20141011192333.GA15412@guitar.localdomain

Re: ISDN call logging

[Mirko Parthey]

On Sun, Jan 05, 2014 at 01:49:30AM +0100, Urs Thuermann wrote:
 My ISP provides VoIP and the DSL modem/router provides an internal
 ISDN S0 bus where my ISDN telephones are attached.  I have also hooked
 an old Linux server with an ISDN card to the S0 bus.  This server runs
 a Linux 2.4.37.8 kernel with isdn4linux and isdnlog to log all
 incoming and outgoing calls, as well as call alerts.
 
 Now I want to retire that old server and do the call logging on
 another server running Debian jessie.  Unfortunately, ISDN seems to be
 poorly maintained nowadays and I wasn't able to do this.  The kernel
 has mISDN instead of isdn4linux while some of the documentation is
 from the Linux 2.4 or even Linux 2.2 era.  The man page isdnlog(8)
 states that isdnlog only works with the HiSax driver i.e. isdn4linux.
 Using the Debian kernel and ISDN utils I couldn't get any call logs.
 
 I compiled a new Linux 3.12.6 kernel with isdn4linux and HiSax and
 tried running isdnlog with it.  This actually worked and logged calls
 like my old Linux 2.4-based server, however, isdnlog crashes after
 some random time (couple of days) with segmentation fault.  I haven't
 yet had the time to debug this problem.
 
 So, is there any other way to get reliable logging of alerts and calls
 to/from my ISDN phones on a current Debian system with ISDN card?

With mISDN superseding isdn4linux, the layer 3 (call setup etc.) is now
completely handled in userspace, and Linux Call Router (LCR) seems to
be the recommended mISDN layer 3 application.  I don't know if it can be
configured for your use case.
http://www.linux-call-router.de/

As an alternative, there is a CAPI emulation on top of mISDN, so if you
find a CAPI logger, this could be your solution.
http://www.isdn4linux.de/pipermail/isdn4linux/2012-January/005580.html

Then there is misdn_log, but it is a low-level debugging tool and does
not match isdnlog's features. In particular, it does not have any
knowledge about layer 3.
https://www.misdn.eu/wiki/Misdn_log/
If you can program in C, you could try to extend it for your purposes,
or you could try to port isdnlog to mISDN.

Support options for misdn are listed at
https://www.misdn.eu/wiki/Support/

Despite the old name, mISDN discussions are on-topic there:
https://www.isdn4linux.de/mailman/listinfo/isdn4linux

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20140109163511.GA4871@guitar.localdomain

Re: mirror from archive.debian.org

[Mirko Parthey]

On Thu, Jul 12, 2012 at 12:00:40PM +0400, Anton Gorlov wrote:
 How I get full  mirror of repository for some distr. example lenny amd 64?
 All distributions are linked to the content in the pool/ directory,
 but we need to mirror only one distribution/architecture.
 
 How?

One option would be debmirror.

In addition to http and ftp, it also supports rsync for file transfers,
and it can mirror a subset of distributions and architectures.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120714101929.GA4219@guitar.localdomain

Re: [Feedback needed] Support for ISDN passive fax cards

[Mirko Parthey]

On Tue, May 08, 2012 at 05:56:49PM +, Camaleón wrote:
 I am considering in adding a couple of PCI based Eicon Diva 2.02 
 passive cards on a system running hylafax server. The fax server is now 
 using a couple of external RS-232 modems that work like a charm but as I 
 have a pair of unused ISDN cards I'm consider in adding them to the 
 current setup to enhance the installation.

Please ask yourself which aspect of your installation you expect to
improve by doing this. Fax with passive ISDN cards will be harder to
set up than with analog modems.

 But now I see that HiSax has been deprecated in favor of mISDN and I 
 wonder what's the current status of this driver in Debian, what kind of 
 devices are supported and what are your average experience in using ISDN 
 passive cards with Hylafax.

Check out this announcement:
http://www.isdn4linux.de/pipermail/isdn4linux/2012-January/005580.html
You need very recent sofware to use it, but this seems to be the way
forward.

You may want to ask on this mailing list yourself. Despite its name, it
mostly deals with misdn nowadays.
https://www.isdn4linux.de/mailman/listinfo/isdn4linux
(you have to subscribe to be able to post)

The documentation on www.misdn.org is a bit outdated, but may still be
partially helpful. Look for misdn v2, and ignore misdn v1.

Sorry, I cannot help you with first-hand experience concerning fax support.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120513130634.GA30714@guitar.localdomain

Re: audio CDs and cdrdao vs. cdparanoia

[Mirko Parthey]

On Mon, Jul 05, 2010 at 10:43:44PM +0200, Urs Thuermann wrote:
 For each CD I run
 
 cdrdao read-cd --datafile data.cdr --device /dev/sg0 toc
 and
 cdparanoia -d /dev/sg0 -B
 
 where /dev/sg0 refers to an Plextor Ultraplex 40max SCSI CDROM drive.
 
 [...]

 I can run cdrdao and cdparanoia
 repeatedly, say 10 times, and I get deterministic results, i.e. all
 runs of cdrdao give the same result and all runs of cdparanoia give
 the same result but the results of cdrdao and cdparanoia differ.
 
 Now my question is where these differences come from and which results
 are the correct (better) ones.

With a suitable CDROM drive such as yours and CDs in reasonably good
condition, there should be no need for cdparanoia's data correction
feature.  In fact, I found that it can sometimes do more harm than
good, so I recommend disabling it (-Z), if only for testing purposes.

You could compare your results to Exact Audio Copy (Windows, free for
non-commercial use), which reportedly also runs on WINE.

Another option would bei morituri, a CD ripper for Linux modelled after
Exact Audio Copy.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/20100706110434.ga20...@titus.informatik.tu-chemnitz.de

Re: [LONG] Trouble using debmirror on Lenny (certain public keys not found)

[Mirko Parthey]

On Thu, Jun 24, 2010 at 11:45:31AM +0200, Holger Rauch wrote:
  gpgv: keyblock resource /root/.gnupg/trustedkeys.gpg':
 general error
 gpgv: Signature made Sa 30 Jan 2010 00:18:35 CET using RSA key ID 55BE302B
 [GNUPG:] ERRSIG 9AA38DCD55BE302B 1 2 00 1264807115 9
 [GNUPG:] NO_PUBKEY 9AA38DCD55BE302B
 gpgv: Can't check signature: public key not found
 
 ===
 
 So, I tried to obtain all keys like this:
 
 ===
 
 gpg --keyring /usr/share/keyrings/debian-role-keys.gpg --export | gpg --import
 gpg --keyring /usr/share/keyrings/debian-keyring.gpg --export | gpg --import
 gpg --keyring /usr/share/keyrings/debian-archive-keyring.gpg --export | gpg 
 --import
 
 ===
 
 Furthermore, I tried to explicitly obtain the keys mentioned in the
 error messages:
 
 ===
 
 gpg --keyserver keyring.debian.org --recv-keys 55BE302B
 gpg: requesting key 55BE302B from hkp server keyring.debian.org
 gpgkeys: key 55BE302B not found on keyserver
 gpg: no valid OpenPGP data found.
 gpg: Total number processed: 0
 
 gpg --keyserver keyring.debian.org --recv-keys F42584E6
 gpg: requesting key F42584E6 from hkp server keyring.debian.org
 gpgkeys: key F42584E6 not found on keyserver
 gpg: no valid OpenPGP data found.
 gpg: Total number processed: 0
 
 ===
 
 === Obviously, the keys are not found.
 
 How am I supposed to proceed from this point?

By default, gpg stores new public keys in ~/.gnupg/pubring.gpg,
while gpgv expects them in ~/.gnupg/trustedkeys.gpg.

The following command should import keys straight into trustedkeys.gpg:
  gpg --no-default-keyring --keyring trustedkeys.gpg --import
At least, that's what the debmirror(1) manpage suggests.

If you want to use the same keys for debmirror and apt,
you can simply set a symlink like this:
/home/mirror/.gnupg/trustedkeys.gpg - /etc/apt/trusted.gpg
It may be a good idea to run debmirror under a separate
user account to avoid interference with other usage of gpg.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/20100624113915.ga4...@titus.informatik.tu-chemnitz.de

Re: bandwidth measurement by process

[Mirko Parthey]

On Sun, May 09, 2010 at 02:08:15AM -0400, Celejar wrote:
 I'm looking for a tool that will measure bandwidth used by a specific
 process. I'd like some sort of utility or wrapper that will answer
 questions like How much data did my network backup job push over the
 wire?

Check out network namespaces:
http://lxc.sourceforge.net/network/configuration.php

Put your process(es) in a separate network namespace,
set up virtual networking with macvlan or veth,
retrieve the desired statistics from inside this namespace.

Haven't tried it myself yet, though.

Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/20100510160631.ga30...@titus.informatik.tu-chemnitz.de

Re: Debian 5 server - auto-reconnect to wireless networks?

[Mirko Parthey]

On Sun, Feb 21, 2010 at 04:46:51PM +, chombee wrote:
 I'm wondering how to get my Debian server to automatically reconnect to
 my wireless network. The server is an OLPC XO-1 running Debian 5 Stable
 via the DebXO distribution. It does not have a wired ethernet port, it
 connects to the internet through my home wireless network. The wireless
 router occasionally gets restarted, sometimes when I'm not home, and
 this requires me to physically go to the laptop to manually reconnect
 either by operating the network manager applet with the mouse or by
 running a script that uses ifconfig, iwconfig and dhclient.

The network manager package description says it is not intended for
usage on servers.  If you want an interface to be permanently up,
without even logging in, you should set it up in
/etc/network/interfaces.

Check out /usr/share/doc/wpasupplicant/README.modes.gz from the
wpasupplicant package. Your case should be covered by section 2:
Mode #1: Managed Mode.

Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/20100222115524.ga8...@titus.informatik.tu-chemnitz.de

Re: sshd

[Mirko Parthey]

On Sat, Oct 31, 2009 at 02:43:09PM +0100, Roy Stuivenberg wrote:
 I have a question concerning sshd.
 I'm unable to connect from outside to my Debian box.
 ...
 (Connection failed Service sshd 
 
 Date:Sat, 31 Oct 2009 03:51:03 +0100
 Action:  restart
 Host:rs-debian
 Description: 'sshd' failed, cannot open a connection to
 INET[localhost:50022] via TCP
 
 
 Regards,
 
 Roy Stuivenberg.
 
 ---
 sshd_config :
 
 # What ports, IPs and protocols we listen for
 Port 22

Looks like your ssh client tries to connect to TCP port 50022, while the
ssh server is listening on port 22 (the standard for ssh).
The non-standard port may be set up on the command line or in a config
file:
$HOME/.ssh/config
/etc/ssh/ssh_config

Check on the server machine if the ssh server is listening:
netstat -ntl | grep 22

If this doesn't help, please post the commands you typed and their
output.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Re: Libata in Debian GNU/Linux Lenny

[Mirko Parthey]

On Sun, Oct 11, 2009 at 12:55:28PM -0300, Daniel Bareiro wrote:
   I'm using Debian GNU/Linux Lenny with linux-image-2.6.26-2-amd64
   stock kernel installed from a binary image from Debian repositories.
  
   In this case the disks are seen like hdX but using kernel 2.6.31.2
   compiled by myself with the sources of kernel.org the disks are seen
   like sdX. I consider that this difference is because with 2.6.31.2
   libata is being used whereas in 2.6.26 it is disabled. Is it
   possible?
 
 With  2.6.31.2 from kernel.org I got the following output with lspci -v:
 
 00:01.1 IDE interface: Intel Corporation 82371SB PIIX3 IDE [Natoma/Triton II] 
 (prog-if 80 [Master])
 [...]
 Kernel driver in use: ata_piix
 Kernel modules: ata_generic, ata_piix, ide-pci-generic, piix
 
 whereas with linux-image-2.6.26-2-amd64 stock kernel installed from a
 binary image from Debian repositories I got the following output with
 lspci -v:
 
 00:01.1 IDE interface: Intel Corporation 82371SB PIIX3 IDE [Natoma/Triton II] 
 (prog-if 80 [Master])
 [...]
 Kernel driver in use: PIIX_IDE
 Kernel modules: ata_generic, ide-pci-generic, piix

Debian's kernel is patched such that ata_piix only handles SATA controllers,
and the old piix driver takes over for PATA controllers:
http://svn.debian.org/viewsvn/kernel/dists/trunk/linux-2.6/debian/patches/debian/drivers-ata-ata_piix-postpone-pata.patch?revision=13847view=markup

I guess this was done while libata was still under heavy development,
and the old driver was considered more stable then.

If you are curious, ask the Debian Kernel Maintainers.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Re: color management?

[Mirko Parthey]

On Sun, Aug 09, 2009 at 09:23:57PM -0700, jeremy jozwik wrote:
 heya list. ive been working with linux on adjusting my photos for the
 last 3 weeks or so. since ive started ive noticed that the way the
 photos look on my linux laptop are CONSIDERABLY different then the
 color that is shown on a windows machine.
 
 now i would not worry too much about this if it were not for the fact
 that every windows machine i see my uploaded photos with shifts the
 colors.
 
 which leads me to believe there is something amiss on my laptop, the
 source of all the color adjustments.
 
 help!
 
 also, i have gimp set to no color management which i found to be the
 setting of the most use on windows when i was using it there.

Laptop screens are optimized for energy efficiency instead of image
quality; their color reproduction may not be as good as an external
monitor and depends strongly on the viewing angle.

Your adjustments correct for the Laptop screen's deficiencies,
but you are breaking your pictures in the process.

no color management is seldom the best choice for photo work.

I would suggest you read at least one of the articles listed here:
http://www.outbackphoto.com/myforum/showthread.php?tid=8
or even get the book by Bruce Fraser.

It is important to calibrate and profile your monitor
and use image viewers/editors which support color management:
http://en.wikipedia.org/wiki/Linux_color_management

As mentioned elsewhere in this thread, photos for the Web should be in
the sRGB color space, but likely yours are already.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Re: Cross-compiler tool chain

[Mirko Parthey]

On Thu, May 28, 2009 at 01:33:26PM +0200, Oliver Schneider wrote:
 Hi,
 
 Debian is available for a whole lot of different architectures. Is there
 a cross compiler tool chain

What are you planning to do?
You may find Emdebian useful:
http://www.emdebian.org/

 which is used to do the builds and if so, is
 it available to the Debian users?

Official Debian packages are built natively on the target architecture.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Re: Does anyone understand terminal job control?

[Mirko Parthey]

On Sat, Feb 07, 2009 at 04:58:13PM -0800, Daniel Burrows wrote:
 directly under my controller process.  It's not even that they're
 starting as foreground processes: I can start them without access to
 the controlling terminal, and they never see a SIGTTIN.

What do you mean by: without access to the controlling terminal?
Is the controlling terminal actually connected to the stdin of your
subprocess in the normal case (check in /proc to be sure)?

The libc info also has example code for implementing a shell with job
control, but I haven't looked at it closely yet.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Re: screen resolution question

[Mirko Parthey]

On Wed, May 21, 2008 at 01:45:44AM +, Glenn Becker wrote:
 I am running a testing box. Recently, it seems, my X settings 
 changed (w/o my intentionally changing them, at least as far as I can 
 recall) so that the screen resolution on Gnome, for example, it was 
 defaulting to 1400x1050. It is nice to see that my older laptop can 
 handle this, but ... I also have older eyes and prefer a setting of 
 1024x768.
 
 So I went looking into the xorg.conf file, and found that ... no default 
 screen res appeared to be set! In fact the file contains almost _no_ 
 details and consists mainly of a lot of sections that look like this:
 
 Section Device
 Identifier  Configured Video Device
 
 ... making it look like a lot of the X configuration is handled 
 automatically.
 
 Well, great ... cough ... I would rather it _wasn't_ handled 
 automatically, because I want to choose my own settings. I've handrolled 
 my own .conf files for X for ages ... is there any way for me to do this 
 and not have my settings overwritten by an automated thingum that has no 
 idea what my specific needs are?
 
 If this is an attempt at an improvement I cannot say I care much for it. 
 Certainly I can change the settings via the Gnome GUI, but this doesn't 
 help me if (as often happens on my old machine) I'd rather just run 
 Fluxbox.

Please see my reply to a similar question at
http://lists.debian.org/debian-user/2008/05/msg01804.html

Best regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Default X resolution

[Mirko Parthey]

On Mon, May 19, 2008 at 11:58:24AM -0400, Damon L. Chesser wrote:
 Is it possible to get x to use 1024x768 by default that way xfree86 used 
 to do? I would still like other resolutions available so I can use them 
 when I need them but most of the time 1024x768 does what I want.
 
 Yes.
 
 man xorg.conf.
 
 Subsection Display
 Depth 24
 Modes 1024x768 some other value here a nother value here
 EndSubSection

Your solution applies to Debian Etch, but it didn't work for me
in testing/unstable anymore.

This page gives more details about modern X configuration:
http://wiki.debian.org/XStrikeForce/HowToRandR12

I added this line to the Monitor Section:
Option PreferredMode 1280x960
It even worked without a modeline - X already came with a suitable mode.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: prevent dpkg from (re-)starting services

[Mirko Parthey]

On Thu, May 08, 2008 at 09:21:35PM +0200, Olaf Leidinger wrote:
 On a file server I installed several debian(-based) distributions into
 an exported directory using debootstrap. These are used as root
 filesystems for diskless clients. To install new packages/updates I
 chroot into the directories of the corresponding distributions and run
 aptitude. The problem is, that every time I install the update of a
 service (e.g. cups) or a new one, the service is started in the chrooted
 environment, but I don't want to run any service in there.
 
 Is there a way to prevent dpkg from (re-)starting these services?
 Deleting them from the /etc/rc* directories is not an option, as they
 are needed by the clients.

Yes, it is possible to conditionally prevent services from
being (re-)started, with the init script policy mechanism.
It's explained in the invoke-rc.d manpage and in the file
/usr/share/doc/sysv-rc/README.policy-rc.d.gz

For a local chroot which is not shared, I use a simple
/usr/sbin/policy-rc.d script just containing exit 101,
but the mechanism should also be flexible enough for your needs.

Best regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

email without From: header (was: disassembling machine code)

[Mirko Parthey]

On Sat, Mar 22, 2008 at 08:33:52PM -0700, Daniel Burrows wrote:
  Mutt may be a lesser piece of crap than most mailers, but it is still a
  piece of crap.  It screws up when From: is missing, and acts as if you
  had send it.
 
   I think it may be worse than that: I set my MTA up to add a default
 From field to outgoing email, but it ended up applying to incoming mail
 as well.
 
   In which case it's not Mirko's fault at all.  Mirko, I'm sorry I
 jumped on you like that.  Probably I should have realized something
 screwy was going on and checked other possibilities first.

No problem. I'll try to explain what happened.

When first replying to Peter Easthope, I got distracted and mistakenly
sent him a personal email without Cc'ing debian-user.
After recognizing my mistake, I retrieved the email from my sent folder
and bounced it to debian-user.

Without going into the details of my email setup, the result was
that I sent a broken email without From: to the list.
Sorry for that, and for the resulting confusion.

I'll thouroughly investigate this message's From: header when I receive
it back from the list. :-)

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: disassembling machine code

[Mirko Parthey]

On Fri, Mar 21, 2008 at 01:39:09PM +0100, [EMAIL PROTECTED] wrote:
 On Wed, Mar 19, 2008 at 12:26:19PM -0700, PETER EASTHOPE wrote:
  I have these 5 bytes of machine code to 
  disassemble.
  
  b8 12 00 cd 10
  
  I've looked at gdb and objdump.  Appears they 
  need a complete object file.  Someone please 
  give a clue.
 
 Write the machine code to a file:
 $ perl -e 'print pack H*, b81200cd10;'  /tmp/binfile
 $ hd /tmp/binfile
   b8 12 00 cd 10|¸..Í.|
 0005
 
 Then use objdump, specifying the object file format as raw.
 Because the raw file has no metadata, you need to give the CPU
 architecture as well.
 
 $ objdump -m i386 -b binary -D /tmp/binfile

Sorry, I meant to write i8086 instead of i386:
$ objdump -m i8086 -b binary -D /tmp/binfile
because because this looks like 16 bit x86 code to me.
You should have told us the CPU architecture in the first place...

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Modem in Debian

[Mirko Parthey]

On Sun, Feb 03, 2008 at 08:16:57PM +, Pantor wrote:
 there is a modem plugged to PCI slot and no idea what to do with it.
 Any suggestions, advices, please.
 
 Regards.
 Andrius

Check out http://linmodems.technion.ac.il/
and run the scanmodem tool provided there -
it should produce useful documentation for your modem.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: DVD drive misbehaves after a while, kernel problem?

[Mirko Parthey]

On Tue, Dec 18, 2007 at 11:14:50PM -0700, Matthew Dale Moore wrote:
  Dec 18 10:28:00 guitar2 kernel: hda: cdrom_pc_intr: The drive appears
  confused (ireason = 0x01). Trying to recover by ending request. Dec 18
  10:28:00 guitar2 kernel: hda: status error: status=0x58 { DriveReady
  SeekComplete DataRequest } Dec 18 10:28:00 guitar2 kernel: ide: failed
  opcode was: unknown Dec 18 10:28:00 guitar2 kernel: hda: drive not ready
  for command Dec 18 10:28:02 guitar2 kernel: hda: status error: status=0x58
  { DriveReady SeekComplete DataRequest } Dec 18 10:28:02 guitar2 kernel:
  ide: failed opcode was: unknown
 
 I am experiencing this same problem since Jan 2007. I have the same 
 motherboard but a different DVD-RW drive. When I first attempted to install 
 linux on my newly built computer, the damned JMicron controller was 
 unsupported, and so DVD drive access was impossible. About 3 weeks later 
 support was added, and I have always figured that this behavior is a symptom 
 of just poor support for this controller. 
 
 A workaround that I have been using for the past few months is just to keep a 
 blank CD-RW (or some kind of media) in the drive at all times. It seems like 
 this only happens when the drive is empty for more than a few minutes.
 
 I would be interested in any advice anyone has to offer.

Thanks for your comment - it encouraged me to dedicate some more time to
this problem, focussing on the jmicron driver.

My system finally stopped printing error messages after applying these steps:
- upgrade the mainboard BIOS to version F12
- blacklist the old IDE driver module for the jmicron controller:
  echo blacklist jmicron  /etc/modprobe.d/_local_
- build kernel 2.6.24-rc6 with CONFIG_PATA_JMICRON=m
  (enables libata support for the jmicron controller)
- reboot and cross fingers :-)

The kernel and BIOS upgrades might not be necessary,
the switch to libata is likely the key step.

To check which driver you are using, you can use:
$ lsmod | grep jmicron
The old module is called jmicron, the new one pata_jmicron.

Also, your DVD drive will change its names from /dev/hdX and /dev/cdrom
to /dev/scd0 and /dev/cdrom1.

If you decide to try this, please post your results.

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

DVD drive misbehaves after a while, kernel problem?

[Mirko Parthey]

After between 10 and 60 minutes of uptime, my DVD drive begins to show
strange behaviour:

- an icon appears on the KDE desktop showing either a data CD-ROM
  or an audio CD (although the tray is empty and closed)

- in /var/log/kern.log, these messages appear in large numbers:

Dec 18 10:28:00 guitar2 kernel: hda: cdrom_pc_intr: The drive appears 
confused (ireason = 0x01). Trying to recover by ending request.
Dec 18 10:28:00 guitar2 kernel: hda: status error: status=0x58 { DriveReady 
SeekComplete DataRequest }
Dec 18 10:28:00 guitar2 kernel: ide: failed opcode was: unknown
Dec 18 10:28:00 guitar2 kernel: hda: drive not ready for command
Dec 18 10:28:02 guitar2 kernel: hda: status error: status=0x58 { DriveReady 
SeekComplete DataRequest }
Dec 18 10:28:02 guitar2 kernel: ide: failed opcode was: unknown

At this stage, the drive becomes unusable, while it it is working fine 
otherwise.
A reboot fixes the problem, until the next occurence.
  
I am not sure where to report this bug - maybe for linux-image-2.6.23-1-amd64?
Has anybody seen something like this before?
Any hints how I could rule out faulty hardware?

I am running debian testing (amd64), with both 64-bit kernel and userspace.
The problem appears with any of the 2.6.18-5 (stable), 2.6.22-3 (testing)
and 2.6.23-1 (unstable) kernel packages.

my hardware:
- PLEXTOR DVDR PX-760A, firmware 1.07
 (restored to original Plextor firmware for this bug report)
- mainboard Gigabyte 965P-S3, with this IDE controller:
  03:00.1 IDE interface: JMicron Technologies, Inc. JMicron 20360/20363 AHCI 
Controller (rev 02)

Any further information I should provide?

Thanks,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Connect to DOS box

[Mirko Parthey]

On Fri, Jun 22, 2007 at 08:58:33PM +0100, John K Masters wrote:
 At work I use a networked box that runs from Novell Netware on DOS v5.0
 This runs very old accounting software. The box itself is a P120 with
 16 MB RAM and a small HD (about 128MB). No OS on the HD.
 
 I have managed to persuade the somewhat ITphobic management to provide me
 a standalone box on which to analyse the daily data. Obviously I have
 installed Debian Etch. Having written a few Python scripts to take the
 output of printed reports and parse them into an OO spreadsheet it is
 all coming together.
 
 However, to transfer the files I have to use a floppy disc. How can I
 connect the 2 boxes? The DOS box cannot boot from CD, has no installed
 OS (boots into network off floppy) and has only a serial connection
 spare. The network is 10BASE-2 thin ethernet.
 
 Regards, John

Another useful software for your toolbox could be etherboot.
A few years ago, I used a floppy image generated by their rom-o-matic to boot
KNOPPIX over the network on an old machine without PXE support.
The servers in KNOPPIX supporting network boot (DHCP, TFTP, ...) are called
terminal server, which is somewhat misleading.

I'm not sure if it is possible at all to run a full-featured Linux live
CD on a machine with 16 MB of RAM, even in console mode without X.
You might want to try an older or cut-down live CD distribution with a
smaller RAM footprint.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Administration (+apt-get dist-upgrade) of 100s of machines

[Mirko Parthey]

On Thu, Apr 19, 2007 at 10:37:54AM -, Peter Valdemar Morch wrote:
 We have 100s of almost identical machines that need to be kept
 up-to-date with apt-get dist-upgrade .
 
 Having to run apt-get dist-upgrade manually on all of them is just not
 working (taking too much man-power) due to having to answer the same Y/N
 debconf(?) questions over and over again especially about config files.
 
 Is there a smarter way? How does one manage many, many debian
 installations without having to give each one special manual treatment?
 Ideally I'd like this to be a fully automated operation and only be 
 notified of any failures in a reliable way.
 
 Rather than trying to write our own scripts, I've searched the
 admin::configuring and use::configuring tags and come up with dpsyco,
 ugrade-system and fai, but they all seem inappropriate to our needs.
 
 We can't possibly be the first to need this...

For about 20 machines, we use cfengine for configuration management and
a Perl rewrite of pkgsync:
http://rtg.informatik.tu-chemnitz.de/~mpa/software/pkgsync
This version of pkgsync is customized for our cfengine setup - if you
consider using it, please read and adapt it first.

We have error notifications sent via email, you could also write them
to logfiles instead.

We point our sources.list to mirrors of debian.org, backports.org,
and an archive of .debs we built ourselves. If you keep local copies of
all repositories, you can check for upgrade problems before rolling out
new packages to large numbers of machines.

best regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: secure-testing.debian.net down?

[Mirko Parthey]

On Mon, Jun 12, 2006 at 02:28:40AM -0400, Andrew Schulman wrote:
  On May 12th, the testing security archive moved:
  http://lists.debian.org/debian-devel-announce/2006/05/msg6.html
 
 sigh Thanks.  I wish there were a better way of distributing notice of
 these sorts of changes.  debian-devel-announce?  Er, sorry, I don't read
 that one.  I do subscribe to debian-security-announce, but there was nothing
 about the change there.  Also I notice that the Debian security pages are
 completely out of date about this-- they still say that there's no security
 support for testing.

Security support for Debian stable and Debian testing is provided by
different teams with separate mailing lists.

The debian-security-announce list is about stable security, while
secure-testing-announce@lists.alioth.debian.org deals with testing
security. The above announcement was actually sent to the latter, in
addition to debian-devel-announce - and it also contains instructions
for subscribing.

Apart from this announcement, however, the secure-testing-announce list
has been fairly quiet in recent months.  The testing security team
appears to follow a policy of wait until fixed packages move into
testing through unstable, rather than feed fixed packages into the
testing security archive by backporting the fix from unstable.
So one might argue that the Debian security FAQ accurately
describes the current state of testing security support.

regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: what happened to /dev/fd0?

[Mirko Parthey]

On Sun, Feb 26, 2006 at 08:05:08PM -0500, [EMAIL PROTECTED] wrote:
 I have a floppy drive on my srge system.  I used to access it as 
 /dev/fd0.  But there's no /dev/fd0 in my sarge system any more.
 There is a /dev/fd/0, /dev/fd/1, /dev/fd/2, and /dev/fd/3.
 
 Could /dev/fd/0 be the new name for /dev/fd/0?  Or do I need to look 
 elsewhere for the problem?

Are you using udev?  Then you need to load the floppy kernel module
before the device nodes for the floppy drive will appear.

/dev/fd/0, /dev/fd/1, etc. are completely unrelated to the floppy disk
driver.  They are alias names for the file handles (open files) managed
by the current process. By convention, '0' is standard input, '1' is
standard output and '2' is standard error output, all of them usually as
inherited from the parent process.
'3' and onwards are file handles which the process itself opened.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Continuing chant hde:hde1

[Mirko Parthey]

On Sun, Feb 19, 2006 at 09:24:32PM -0500, John wrote:
 My problem continues: an endlessly repeated chant of hde:hde1
 interspersed with an occasional hdb:hdb4. I can't figure out the
 cause, and would be grateful for insight or a lead.

If this disk is attached via cardbus, it counts as removable, and
the following bug report might apply:
  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=350235

Please add your comments there, if applicable.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Clock jumps forward.

[Mirko Parthey]

On Sat, Feb 18, 2006 at 10:53:09PM +0200, David Baron wrote:
 Ntpdate is what I am using. Maybe the setup needs be changed.

Which timeservers are you getting the time from? (see /etc/default/ntpdate)
Do they provide the correct time?

Try /etc/init.d/ntpdate start, and look in /var/log/syslog for new
entries from ntpdate.

If you are using pool.ntp.org, a server is randomly chosen from the pool.
You might want to configure ntpdate to always use the same server(s) -
this makes it easier to analyze the problem.

For debugging, you can also use the command ntdpate -b -u SERVER, so
the output is sent to your terminal rather than syslog.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Clock jumps forward.

[Mirko Parthey]

On Thu, Feb 16, 2006 at 05:27:50PM +0200, David Baron wrote:
 Every few days, I find my clock two hours fast. Easy enough to reset but ... 
 why?
 
 I am running 2.6.15 kernel, Sid, and time is updated using ntp. Since my time 
 zone is universal + two hours, maybe the two hours means something. Bug?

Does this happen after a reboot, and is your /usr filesystem separate
from the root filesystem?  If so, you can find an explanation in Debian
bug #342887 (search for Henrique).

As a solution, you can replace the /etc/localtime symlink by a copy of
the actual file it points to. However, this has the consequence that you
will not benefit from future updates to the timezone files.

You did not specify which ntp tool you are using. ntpd refuses to
adjust your time if it is too far off, so in addition you might want to
use ntpdate, which adjusts the time in a single step during bootup.
It could also work as a band-aid for setting your time correctly
if you do not want to apply the change to /etc/localtime.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Install from usb-stick

[Mirko Parthey]

On Thu, Dec 08, 2005 at 12:49:39PM +0100, thierry wrote:
 Installing from a usb-stick using either of the 2 way on this link:
 http://www.debian.org/releases/sarge/i386/ch04s04.html
 I have the following problem:
 Boots is OK, I can choose language, keyboard, but then the netinst.iso 
 which is on the stick is not founded by the installer.
 I tried on 2 different boxes, with either stable or testing netinst iso.
 Can you give me a hint on what's going on.
 Thierry

I have no idea why this doesn't work for you, but may I suggest a
workaround:
If you have access to a Debian mirror, you could use the netboot initrd,
so you don't need the *.iso anymore.

best regards
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: newer kernel for non-bootable sarge

[Mirko Parthey]

On Fri, Nov 18, 2005 at 08:42:48PM -0700, David Emerson wrote:
 I ran into a nasty problem while trying to install sarge: The 2.6.8
 kernel has a bug that won't allow my computer to boot.
 
 ((  I have a compaq armada 1750 laptop. It has issues... ))
 ((  In the past, I've never been able to get my mouse to work with a
 2.4 kernel, and it's always been easy with 2.6. ))
 
 So, is there a way to download a newer (like, 2.6.12) kernel and just
 drop it into /boot ?? (Keeping in mind that I can't actually boot
 debian to install it)
 
 Where on earth can I download the files? Assumably, the files I want are:
 
 initrd.img-2.6.12-10-386
 vmlinuz-2.6.12-10-386
 
 And probably also:
 
 config-2.6.12-10-386
 System.map-2.6.12-10-386
 
 Assuming I can find the files to download, should this operation be as
 simple as I'm assuming it to be?

With the installer, you cannot just download a new kernel, drop it into
some directory, and hope that it works. You would have to rebuild the
installer to replace the kernel. To succeed, you need experience with
Debian packaging. I'd suggest to only try that after everything else
has failed.

If the sarge installer works with kernel 2.4, you could always install
using 2.4 and upgrade the kernel to 2.6 after the installation, to make
your mouse work.

If you already need kernel 2.6 during installation, there are still
simpler alternatives than rebulding the sarge installer:

- use the etch beta 1 of the Debian installer
  see: http://www.debian.org/devel/debian-installer/
  It has kernel 2.6.12 and should also be able to install sarge.
  If it works for you, this should not be harder than using the sarge
  installer directly.

or
- boot from a live CD such as Knoppix or Kanotix,
  and follow the instructions given here:
  http://www.inittab.de/manuals/debootstrap.html
  Again, some experience with Unix/Linux is necessary.

best regards
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: kernel upgrade questions

[Mirko Parthey]

On Wed, Nov 23, 2005 at 05:35:49PM -0500, David Zelinsky wrote:
 I'm trying to upgrade the kernel from 2.4 to 2.6 on my Thinkpad.  I'm
 using custom built kernel rather than the stock debian kernel, since
 the stock kernel doesn't have some features that I want.
 
 My question is how to deal with all entries in /etc/modules.conf and
 /etc/modules that refer to modules that don't exist in 2.6, and the
 new modules which didn't exist in 2.4.  I would like to be able to
 boot into the old 2.4 kernel and have everything work as before.  Is
 there a way to have entries in these files be conditional on the
 kernel version?

Please see the /etc/init.d/module-init-tools script on how to make sure
/etc/modules is chosen depending on the currently running kernel.
Naming the files /etc/modules-2.4 and /etc/modules-2.6 should work.
You will rarely need to load any modules explicitely if you have
discover/hotplug or something similar installed.

Under kernel 2.6, modprobe does not use /etc/modules.conf anymore.
Have a look at the files in the /etc/modprobe.d/ directory instead.

best regards
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: [Q] partman, big disk and old BIOS : the whole capacity is not seen

[Mirko Parthey]

On Thu, Nov 24, 2005 at 08:22:38AM +0100, [EMAIL PROTECTED] wrote:
 Using a utility executable provided by the manufacturer of the disk, I trick 
 the BIOS which now believe that the disk capacity is around 32 GO.
 
 The problem is that partman in the net-install CD of Sarge is also tricked, 
 and only saw the 32 GO.

Please add a hda=stroke parameter when booting the installer.
Check the installer's online help on how to do this.
(Replace hda with the actual name of your harddisk, if it is different).

You can find further explanations in the Large-Disk-HOWTO.
In particular:
http://www.tldp.org/HOWTO/Large-Disk-HOWTO-11.html#ss11.6

best regards
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Want to move from root LVM/LILO to LVM/Grub

[Mirko Parthey]

On Mon, Oct 03, 2005 at 01:06:16PM -0500, Jason Martens wrote:
 I believe your /boot partition needs to be on a non LVM partition.
 Well, I can boot just fine with a root partition on the LVM volume using
 LILO.  I believe this is because my initrd has the LVM drivers included,
 so it can recognize the root partition.
 he said /boot, not 'root'.
 Ah, so he did.  However, as I said earlier, I can boot even though I 
 have no /boot partition, and root is in LVM.

There is a difference between lilo and grub in how they find the disk
sectors containing the kernel image and the initial ram disk.

Lilo determines the sector numbers while linux is running (before the
reboot), when you issue the command lilo. It stores these numbers in a
map file, the sector numbers of which are in turn determined and stored
in a reserved region of the lilo boot sector.
On bootup, lilo simply loads the predetermined sectors - first the map
file, then the kernel image and optionally an initial ram disk.
This way, lilo does not need any special support for LVM, it can just
make use of the underlying OS drivers to find the sectors.

Grub tries to find the sectors containing the kernel image during the
boot process. To that end, it has a built-in driver for the root
partition's filesystem.  When installing grub as boot loader, a suitable
variant of grub with the correct filesystem driver was chosen (this
explanation is somewhat simplified). In order to load a kernel image
from a filesystem stored on an LVM volume, grub would need to support
LVM read access in addition to the filesystem.
I didn't verify it, but I suspect there is no grub version with such
support.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: disable ethernet auto-negotiation?

[Mirko Parthey]

On Sat, Aug 27, 2005 at 10:01:15PM -0500, Matthew Lenz wrote:
 Is there a /etc/network/interfaces option to set ethernet speed/duplex and 
 disable auto-negotiation? 

You can use the post-up interface option to run a tool for setting
the speed/duplex. Which tool? That depends. ;-)

Which type of card are you using, and which driver module?

One or more of the following tools may be able to help you:
mii-tool
mii-diag
ethtool
nictools-pci
nictools-nopci

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: How to automatic restore alsamixer settings with alsa-utils?

[Mirko Parthey]

On Fri, Aug 26, 2005 at 09:54:45AM +0200, Tim Ruehsen wrote:
 on my SID system, I experience that alsamixer settings are not restored after 
 booting. It looks like /etc/rcS.d/S50alsa-utils is called at the wrong time 
 (too early?). Because after rebooting I can restore alsamixer settings just 
 by calling '/etc/init.d/alsa-utils start' or alternatively by calling 
 'alsactl restore'.
 
 And yes, alsa-utils is called while booting (I can see it's output).

I experienced the same problem. When alsa-utils was called, the driver
module for my sound card was not loaded yet. alsa-utils wouldn't trigger
autoloading of the module either, so it failed to restore the mixer
settings.  The module was autoloaded later, probably when KDE started up.

My solution was to load the module explicitly in /etc/modules, so it
gets loaded before /etc/init.d/alsa-utils is called.  Installing
discover or hotplug might also help, although I haven't tried this.

There is some documentation at the top of /etc/init.d/alsa, and also in
/usr/share/doc/alsa*/ .

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: How to automatic restore alsamixer settings with alsa-utils?

[Mirko Parthey]

On Fri, Aug 26, 2005 at 12:30:11PM +0200, Tim Ruehsen wrote:
  There is some documentation at the top of /etc/init.d/alsa, and also in
  /usr/share/doc/alsa*/ .
 
 Due to the docs in /etc/init.d/alsa-utils, /etc/rcS.d/S50alsa-utils should be 
 the right choice to restore alsa settings.

What happens when you execute the following?
/etc/init.d/alsa-utils stop
/etc/init.d/alsa-utils start

You could also try unloading and reloading the driver module for your
soundcard in between.
If nothing helps, you can always execute the script with debugging turned on
(set -x) or add some debugging output to it.

 Do you know how to switch /var/log/boot.log (or similar) on? I definitely do 
 not have the boot messages written into a log file. See this:
 -rw-r--r--  1 root root 4036 Dec 29  2003 boot
 -rw-r--r--  1 root root 3306 Dec 11  2003 boot.0
 It stopped ~Dec 29th 2003.
 Some experiments with /etc/syslog.conf and local7.* just generated an empty 
 file.

Set BOOTLOGD_ENABLE=yes in /etc/default/bootlogd.

The bootlog mechanism doesn't use syslog - bootlogd writes to the
logfile directly.

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Printer localhost:631

[Mirko Parthey]

On Thu, Aug 18, 2005 at 09:53:01AM +1000, Marie-Louise wrote:
 This is the message that is in /var/log/cups/error_log
 
 I [18/Aug/2005:08:37:47 +1000] Listening to 0:631
 [...]
 E [18/Aug/2005:08:37:49 +1000] StartListening: Unable to find IP address
 for server name ariestao - Host name lookup failure

The last message seems to be the root of the problem.  Your machine has
a host name of ariestao, but the resolver doesn't know how to translate
this into an IP address. Try the following command:
ping -c1 ariestao

If this produces an error message like
ping: unknown host ariestao, 
the easiest way to fix this would be to add a line to /etc/hosts with
the hosts IP address and the hostname:
192.168.xxx.yyy  ariestao
(adjust 192.168.xxx.yyy to match your actual IP)
Check again with ping afterwards.

If you get a different error, and cannot interpret it, please post it
here.

 
 # telnet localhost 631
 Trying 127.0.0.1...
 telnet: Unable to connect to remote host: Connection timed out

That seems like your networking setup is broken, or you have a firewall
which drops packages. If networking was OK, and no server process was
listening on port 631, you would get the following error:
telnet: Unable to connect to remote host: Connection refused

 
 # /sbin/fuser -n tcp 631
 -su: /sbin/fuser: No such file or directory

The error message gives you a hint why the command failed: On Debian
systems, the fuser command is not in /sbin/fuser, but in /bin/fuser.

Try the following command instead (as root):
fuser -n tcp 631

Or better yet (also as root):
netstat -atnp | grep 631

 Its not a firewall problem, because nothing changes even when there is no
 firewall or I use the command :- # iptables -nL INPUT

Does the command iptables-save output anything?

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Sarge ISA NE2000

[Mirko Parthey]

On Wed, Aug 10, 2005 at 03:42:53PM +1000, Chris Phillips wrote:
 Its worth giving a go, but I still don't understand
 why the options I have set in modules.conf are not
 being picked up.

Are you using Kernel 2.6?
If so, the configuration should be located in
/etc/modprobe.conf or /etc/modprobe.d/ .

See the manpage modprobe(8) for details.
You might want to run modprobe under strace to double-check this.

I have found that recent 2.4 and 2.6 kernels handle ISA PNP devices
themselves, so isapnptools should be unnecessary.
FWIW, I have seen ISA NICs that could be configured by software instead
of jumpers, but they were not ISA PNP compliant - you need to use tools
such as atlantic-setup from the nictools-nopci package to set up their
I/O base address and IRQ and IRQ.

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Can't browse certain websites

[Mirko Parthey]

On Tue, Aug 09, 2005 at 09:30:50PM -0400, Michael D. Crawford wrote:
 I have a complex home LAN setup that is gatewayed to the internet by a 
 Sarge IP Masquerading box.  Most of my machines see the web just fine, 
 but an OS X Mac that is connected via IP over FireWire routed through 
 another OS X Mac using Apple's Internet Sharing can see some websites 
 but not others.  DNS works fine, no error is given, the browser just 
 sits there waiting with no response.

One possible reason is a broken firewall on the server side which blocks
all ICMP messages, including those used for Path MTU discovery
(fragmentation needed, but DF bit set).
I have no experience with ethernet over firewire, but if it uses an MTU
smaller than 1500, this misconfiguration on the server side will cause
you problems. You could of course deactivate PMTU discovery locally,
but if it is active on the server side, together with this strange
firewall setup, all you can do is complain to the server operator.

If you can live with just web access and without full IP connectivity,
I would suggest you set up a proxy on a machine that doesn't show these
problems, and configure your Mac client to use this proxy.

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: proftpd and sshd delays

[Mirko Parthey]

On Fri, Aug 05, 2005 at 03:13:01PM +0800, James Duff wrote:
 Both sshd, and proftpd are very slow to respond. I've noticed this in
 sshd under debian on previous versions and machines also. Sshd asks
 for a username almost immediately when I connect, but there is a
 substantial time before it asks for a password.
 
 Proftpd also has a significant delay before responding to login information.

Is there a packet filtering firewall or a NAT device between your client
machines and the server, or do your clients run a personal firewall
software which drops incoming connections? Or does your server machine
drop outgoing connections?

Your FTP and SSH servers might be configured to ask the client machine
which user issued the connection request. The ident protocol is used
for this purpose (sometimes also called auth, which is a bit misnamed).
If your client machines ignore such connection attempts, instead of
rejecting them, this will cause the delays you observed.

You should be able to analyse this by running a packet sniffer such as
ethereal on the server side.

Another reason could be that the user lookup on the server itself takes
too long. Which user directory are you using - passwd, NIS, LDAP or
something else? If you don't know, please look in /etc/passwd for lines
starting with + (which would indicate you are using NIS), and post
your /etc/nsswitch.conf.

The strace tool, attached to a server process, might help diagnosing
this case.

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: problem with a device 2TB

[Mirko Parthey]

On Thu, Jul 28, 2005 at 12:52:18PM +0200, christian gattermair wrote:
 problem solved, with parted
 thx for help :-)

Now you've made me curious - what did you change since your last email
so that parted now suddenly works?

Also, if you can spare the time, could you please file bug reports with
all the tools that could not handle your large device properly
(e.g. fdisk, cfdisk, blockdev)?

Thanks,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Optimize iso images, find duplicates

[Mirko Parthey]

On Thu, Jul 28, 2005 at 12:57:02AM +0200, Josep M. wrote:
 I would like ask what package can I use for optimize an iso image, I
 explain, when you have a CD with several files with identical content but
 different name the program detect this and only copy one file in the iso
 image and the rest of is a link inside to the iso image, this generate a
 more small iso file when several identical files are present.

The Knoppix remastering scripts by Fabian Franz include a tool to find
identical files and replace one of them with a hardlink to the other.
When creating the iso image with mkisofs, the hardlinked files' contents
will be shared, so they only occupy as much space as one instance.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: use knoppix to rerun lilo

[Mirko Parthey]

On Tue, Jul 26, 2005 at 09:28:19PM +0200, Felix Natter wrote:
 Anyway, usually I can easily fix this by booting Knoppix (3.1 in my case),
 mounting the root filesystem /dev/hda2 (it's all on one partition),
 and running /sbin/lilo from a chroot-shell (chroot /mnt/hda2 /bin/bash).
 But when I do this, I get the error:
 Fatal: open /dev/hda: Permission denied

When you are working in the chroot, /dev/hda is equivalent to
/mnt/hda2/dev/hda as seen from the Knoppix root filesystem.

The above error is probably caused by /mnt/hda2 being mounted with the
option nodev, which is the default in Knoppix. nodev means that
character and block device nodes lose their special meaning and cannot
be use to access the underlying device drivers anymore.

As a solution, don't use the desktop icons to mount the file system, but
use the command line instead, in order to mount it without any options:
  mount /dev/hda2 /mnt/hda2
Of course, you should unmount it first if it was still mounted.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: problem with a device 2TB

[Mirko Parthey]

On Tue, Jul 26, 2005 at 09:52:09AM +0200, christian gattermair wrote:
 i use debian sarge with 2.6.8-2-686-smp and like to use an raid device (sda) 
 with 3 TB.
 
 i have read in the internet and found that there is an kernel feature called 
 LBD. it is activated
 
 but with cfdisk:
 
 cfdisk /dev/sda
 FATAL ERROR: Cannot get disk size
 [...]

Just a guess - have you tried partitioning with GNU parted?
Does blockdev --getsize /dev/sda report the correct number of sectors?

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: audigy 2 value

[Mirko Parthey]

On Thu, Jul 21, 2005 at 11:36:54AM +0100, Keir Lawson wrote:
 I have been trying to set up my newly aquired audigy 2 value on debian
 unstable... but to no avail. alsaconf does not detect it, and i have
 tried to follow all ideas presented by google.  here is the relevant
 line from lspci:
 :00:0b.0 Multimedia audio controller: Creative Labs: Unknown device
 0008
 
 Can anybody help with this?

You might want to try kernel 2.6, if you haven't done yet.

The following commands (run as root) could help you diagnosing the problem.
If you need further help, please post their output:
  modprobe snd_emu10k1
  lsmod | grep emu
  lspci -vv -s 00:0b.0
  lspci -vvn -s 00:0b.0

I have an Audigy 2 here, which is running fine with kernel 2.6.12.
(pci id: 1102:0004, subsystem: 1102:1002)

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Looking for software Debian Sarge

[Mirko Parthey]

On Wed, Jun 29, 2005 at 06:52:28PM -0500, Sergio Cuéllar Valdés wrote:
 Yes, that is a good idea. I have made a couple of debian packages for
 myself. But, how do I look for a sponsor ?  The packages that I 've
 made are very simple, this one, zoneminder, needs a previous
 configuration before compiling it, it also needs apache and mysql, so,
 I think it's more diffcult, isnt't it ?

A good place to ask these questions is the debian-mentors list:
http://lists.debian.org/debian-mentors/

You might want to read the FAQ before posting:
http://people.debian.org/~mpalmer/debian-mentors_FAQ.html

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Software Raid Partition'd?

[Mirko Parthey]

On Thu, Jun 02, 2005 at 08:04:30AM -0700, [EMAIL PROTECTED] wrote:
 I'm configuring a new Sarge box with Sofware Raid 5 using
 4 Hard disks.
 I'm wondering what is the best way to partition the array?
 For example:
 
 Is it best to just create 1 Raid 5 mounted on /
 or
 Create several Raid 5 arrays and split them into
 /
 /home
 /var
 /usr

You might want to add a separate filesystem for /tmp.

 Note: Will use a Raid 0 mounted on /boot
 
 This server will be used with about 200 workstations serving file, web and
 email.
 I guess I'm basically asking if it makes a performance boost to parition
 the drives up, or just have one mega drive mounted on /

In addition to your Raid 0 for /boot, you should only create one Raid 5
array and use LVM to allocate space out of it to your filesystems.
This way, you can conveniently resize the filesystems if necessary.
Having / on LVM also works, although in this case you should prepare a
rescue CD with LVM support (Kanotix would do the job).

I would recommend against putting everything in / - this might be OK on
a machine with a small disk, such as an old notebook, but not on a
server, where a fsck would take ages on such a monster filesystem.

With several Raid 5 arrays, you would need a complete backup/restore
cycle to resize any filesystem.

regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

how to use the LANGUAGE environment variable

[Mirko Parthey]

According to the glibc reference manual, LANGUAGE can contain a
colon-separated list of locales to indicate their order of preference.

However, the behaviour I observed is different from what I expected.
(tested with cat -h, which produces a short error message)
LANGUAGE seems to have no effect when I leave LANG unset,
so I did the experiment with LANG=de_DE and LANG=en_GB.

  +-+-+
  LANGUAGE  LANG | de_DE   | en_GB   |
 V| | |
+-+-+-+
| en_GB   | english | english |
| de_DE   | german  | german  |
| en_GB:de_DE | german  | german  |  - this seems strange to me
| de_DE:en_GB | german  | german  |
+-+-+-+

With LANGUAGE=en_GB:de_DE, I would have expected English output
rather than German.

Now to my question:
Is this a misunderstanding on my part about how LANGUAGE works,
or should I report this as a bug in glibc/locales?

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: lvm2 (sid) and pvmove - device-mapper targets

[Mirko Parthey]

On Mon, May 23, 2005 at 08:50:17PM +0200, Chris Searle wrote:
 But - pvmove gives:
 
 # pvmove /dev/hda7 /dev/hda5
   mirror: Required device-mapper target(s) not detected in your kernel

please try:
modprobe dm-mirror

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Freeswan and kernel 2.4.27

[Mirko Parthey]

On Thu, May 19, 2005 at 06:21:22PM -0500, Jacob S wrote:
 Is anyone using Freeswan with a 2.4.27 kernel? If not, what 2.4 kernel
 are people using?
 
 I've been trying to patch kernel-source-2.4.27 using
 kernel-patch-freeswan and freeswan-modules-source, but keep failing
 while applying the patches. I'm using the provided debian/rules in
 freeswan-modules-source and make-kpkg --added-patches=freeswan for
 kernel-patch-freeswan. I'm doing a make clean and untarring a
 clean copy of kernel-source-2.4.27.tar.bz2 between attempts, but can't
 get either patch to work. And I know I only need one of those patches to
 work - that's all I'm trying for, but they both fail on me.

The Debian kernel is already patched to include a backport of the IPsec
implementation from Linux 2.6. Although I havent't tried this myself,
I would expect this to be the reason why you were not able to apply the
freeswan patch.

Your choices are probably:
- use the Linux 2.6 IPsec, either in 2.4 (backported) or 2.6;
  no further patches need to be applied to the kernel,
  the userspace applications are a different story (ask again if you
  need help there)
- download a vanilla kernel from kernel.org and apply the freeswan
  patch, or any of the derivatives which are still maintained (openswan
  or strongswan, last time I checked)

I hope this helps,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Freeswan and kernel 2.4.27

[Mirko Parthey]

On Fri, May 20, 2005 at 07:13:47AM -0500, Jacob S wrote:
  The Debian kernel is already patched to include a backport of the
  IPsec implementation from Linux 2.6. Although I havent't tried this
  myself, I would expect this to be the reason why you were not able to
  apply the freeswan patch.
 
 I had read that Debian kernels included a backport of IPsec from 2.6 and
 originally tried setting up freeswan without it. However, ipsec barf
 reports that there is no kernel support present.

You need patched freeswan userspace tools in order to use them with the
Linux 2.6 native IPsec or its backport. According to
/usr/share/doc/freeswan/README.Debian.gz, the patch has been included in
the Debian freeswan package since version 2.01-2.
You also need to install ipsec-tools:
http://packages.debian.org/testing/net/ipsec-tools

Please be aware that my experience with freeswan is a bit dated,
all of the above is taken from the documentation or mailinglists.

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: folder display order in mutt

[Mirko Parthey]

On Sat, May 07, 2005 at 05:54:27PM -0700, Todd A. Jacobs wrote:
  locale setting? (LC_COLLATE)
 
 That's it! The good box has everything set to POSIX, while the bad
 box is set to en_US. So, export LC_ALL=POSIX solves the problem.
 
 However, I'm trying to figure out why the boxes are different, since
 they're both running Debian Sarge. Both have the same set of locales
 installed, and neither user has LC* defined in a profile script; where
 is this being set as the system default?

See /etc/environment, which was generated by base-config, I guess.

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: debootstrapping via knoppix, and what next?

[Mirko Parthey]

On Wed, Jan 05, 2005 at 06:08:32AM -0500, David T-G wrote:
 PS -- Does anyone have subscription info for a knoppix-installer list?

I haven't heard of a knoppix-installer list yet. You might want to try
the general Knoppix mailing list:
  [EMAIL PROTECTED]

The subscription front end used to be at
  http://www.linuxtag.org/wws/info/debian-knoppix
but seems to have disappeared now.

Also check out the forums at knoppix.net:
  http://www.knoppix.net/forum/
They have a section on HDD install.

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

removing the CD while running Knoppix

[Mirko Parthey]

On Mon, Dec 20, 2004 at 12:36:04PM -0800, Alvin Oga wrote:
 just look at knoppix as what is symlink to memory
 or turn off that feature or any other cdrom based distro that works
 properly .. ( most don't .. ie.. i want a cd distro that allows 
 the cd to be popped out so that other apps can be loaded )

If you have at least 1GB of RAM, you can boot Knoppix with the option
toram, this will allow you to remove the CD.

To do this with less RAM, you would need to shrink the CD size by
remastering it with some applications removed.

Or you can use the option tohd=/dev/hda1 (adjust device as needed)
if you don't mind touching the contents of your harddisk.

Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Knoppix Mailing List (was: knoppix HD-install problem)

[Mirko Parthey]

On Tue, Dec 14, 2004 at 09:48:50AM -0800, Alvin Oga wrote:
 btw.. knoppix does NOT have a mailing list on their site(s)
   - just a silly forum ( aka messy/unreadable/searchable forums )
 [...]
 since debian's kids ( knppix ) does not have an active mailing list,
 it'd be good to provide a safe haven ( safe mailing list ) for the
 grandkids

The mailinglist is active (again).
- subscription/archive:
http://www.linuxtag.org/wws/info/debian-knoppix
- to post:
[EMAIL PROTECTED]

You can normally find the link at knoppix.net.  Unfortunately, their
documentation section is currently unavailable because of a wiki
software upgrade.

best regards
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Debian's Linux Kernels: Which patches are applied?

[Mirko Parthey]

On Thu, Nov 18, 2004 at 01:57:24PM +0100, Lukas Ruf wrote:
  how can I find out which patches were applied to kernels as
  distributed with Debian?
  
  Quick answer:
  you can get the information you whant in the distributed tarball
  /usr/src/kernel-source-2.x.x.tar.bz2
 
  For example, in the /usr/src/kernel-source-2.6.9
  there are files with explicit name which may contain what you want:
  Debian.src.changelog
  applied_patches
 
 I see the changelog but I do not see the applied_patches .
 However, I found the packages 'kernel-patch-version'
 
 uccellina:src!85 dpkg -l | grep kernel | grep patches
 ii  kernel-patch-d 2.4.24-3   Debian patches to Linux 2.4.24
 ii  kernel-patch-d 2.6.7-4Debian patches to Linux 2.6.7
 ii  kernel-patch-d 2.6.8-8Debian patches to Linux 2.6.8
 
 What I would like really, is something like applied_patches that would
 explicitely name the patches, i.e. file names of the patches applied.

1. install the package(s) kernel-patch-version for the version(s) you need

2. look in the files
 /usr/src/kernel-patches/all/version/debian/series/*
   for a list of patches that are applied to get the corresponding Debian
   kernel tree;
 + means patch will be applied,
 - means patch will be applied in reverse
 X has special meaning: the named file will be deleted when
 building official Debian kernel-source packages

3. the patches themselves are located in
/usr/src/kernel-patches/all/version/debian/*

4. if you have an unpacked kernel tree, you can switch it forward and
   back between the revisions of the Debian kernel tree as given by the
   sequence files, by running
 /usr/src/kernel-patches/all/version/apply/debian ktver
   where version is the upstream kernel version and ktver is the
   Debian kernel tree revision you would like to reproduce.

These steps should apply to at least 2.4.27, though they may also be
valid for other versions.

Hope this helps,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: Broken Galeon in Sarge

[Mirko Parthey]

On Fri, Nov 21, 2003 at 05:30:25PM -0700, Thanasis Kinias wrote:
 Galeons.  Galeon refuses to execute, claiming that it can't find a gconf
 schema.  I followed the instructions in the Galeon FAQ, but there are no
 gconfd processes running, nor any lockfiles to delete.  It seems that
 the gconf2 I pulled from testing yesterday broke Galeon.

For a workaround, see bug #216430:
quote
  reinstalling schema doesn't work. But
  cd ~/.gconfd
  ln -s /tmp/gconfd-petr/lock .
  solves this problem. This seems like some very strange incompatibility to me.
/quote

On my machine, this made the stable galeon 1.2.x work again,
and also fixed some gconf problems with metacity and themes.

Regards,
Mirko


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Re: mutt: how to change new mail checking frequency (solved)

[Mirko Parthey]

On Sat, Mar 10, 2001 at 10:54:38PM +0100, Mirko Parthey wrote:
 mail_check doesn't have the effect I would expect:
 When a message has been delivered to /var/mail/mpa, mutt takes about
 10 minutes to notice the new mail, despite mail_check being set to small
 values such as 5 or 1 (seconds).
 Pressing ^L makes mutt notice new mail immediately, but I'd rather like
 to automate the process. ;-)

In case anyone cares:
The timeout variable does the job, mail_check doesn't.

Regards,
Mirko

mutt: how to change new mail checking frequency

[Mirko Parthey]

Hi,
mail_check doesn't have the effect I would expect:
When a message has been delivered to /var/mail/mpa, mutt takes about
10 minutes to notice the new mail, despite mail_check being set to small
values such as 5 or 1 (seconds).
Pressing ^L makes mutt notice new mail immediately, but I'd rather like
to automate the process. ;-)

Distribution: stable with all proposed and security updates applied

Package versions:
mutt 1.2.5-4
sendmail 8.9.3-23
fetchmail-ssl 5.5.3-1 (recompiled with Kerberos support)

Any ideas?

Regards,
Mirko

Re: Where are clear-text passwords stored?

[Mirko Parthey]

On Sat, Mar 10, 2001 at 01:23:41PM -0800, kmself@ix.netcom.com wrote:
  Is this list an exhaustive list of places where my dial-up password is
  stored?  I'm about to return a laptop after installing Debian for a
  friend, and want to make sure all my passwords are deleted.
--
http://www.lafn.org/~cymbala/Debian/t4700ct.html#passwords
 
 Authoritative answer:
 
 $ find / -type f -print0 | xargs -0 grep -l $password

You might want to run this as root to ensure you don't miss any files
ordinary users can't read.

Notice that you will likely end up with another instance of $password
in your shell history file, which isn't even reported by this command
because it will only be written on logout.

Workaround: Disable the history or use
  kill -9 $$
to log out.

Regards,
Mirko

mutt's pager and 8 bit chars

[Mirko Parthey]

Hi,
I can't figure out how to make mutt's builtin pager
display characters = chr(128), e.g. German Umlauts (ÄÖÜßäöü),
they are shown as question marks.

Mutt's charset option is set to iso-8859-1. For the headers,
you can take this message as reference.

After setting the pager option to less, these characters are being
displayed just fine, but I don't really want to use an external pager.
Pine also has no problems displaying the message.

Am I missing something obvious?

Thanks for any help,
Mirko

Re: mutt's pager and 8 bit chars (solved)

[Mirko Parthey]

On Wed, Nov 01, 2000 at 10:07:54AM -0500, Chris Gray wrote:
 
  I can't figure out how to make mutt's builtin pager
  display characters = chr(128), e.g. German Umlauts (ÄÖÜßäöü),
  they are shown as question marks.
 
 I have my LANG environment variable set to en_US and umlauts show up
 just fine.  You might try setting yours to whatever is appropriate for
 you.  I guess it would be something like de_DE.

Yes, that works. (both en_US and de_DE)

Thanks a lot!
Mirko

Re: _X11TransSocketUNIXConnect: Can't connect: errno = 111

[Mirko Parthey]

On Thu, 17 Aug 2000, Christophe TROESTLER wrote:

 It's Ok -- I'm not permanently conncted to the net...  But however,
 how do you disable all X connections to my box (from outside, not
 another local user)?

Block tcp connections to port 6000 (for Display :0), and to subsequent
ports (only if you are using Display :1 and above - you don't seem to).

 Some names are really old names for this
 machines that are no longer in use.  Does it matter that they are
 listed?

Nobody can use these keys to connect to X servers that aren't running
anymore. ;-) I don't think it matters.

 Well, these are the official X packages (of potato test 1, I have to
 ugrade to potato final still).

I suggest you try that.

 Well, the key are a subset of those given by xauth list (the ones I am
 currently using I guess).
  ^^^
I'll try to clarify this. There are two ways to connect to your X
server:

1. Unix sockets
- /tmp/.X11-unix/X0 (no portnumber)
- key in .Xauthority like aiglenoir/unix:0
- DISPLAY=:0 (no hostname)
- these only work within local machine

2. TCP sockets
- port 6000 ff.
- key in .Xauthority like aiglenoir:0
- DISPLAY=aiglenoir:0 (contains hostname)
- these work locally and over the net

In order for both to work, you need entries in your .Xauthority both for
unix and TCP sockets, using the same key (the one X is currently using).
In your .Xauthority (displayed using xauth list), this doesn't seem to
be the case - the keys are all different. Check these keys again while
DISPLAY=:0 is still working.
Also, XDM-AUTHORIZATION-1 and MIT-MAGIC-COOKIE-1 keys are not
interchangeable. The above paragraph probably applies twice,
once for each kind of keys. I'm not sure about this -- slink only
uses MIT-MAGIC-COOKIE-1.

Mirko

Re: _X11TransSocketUNIXConnect: Can't connect: errno = 111

[Mirko Parthey]

On Wed, 16 Aug 2000, Christophe TROESTLER wrote:

 I've got a funny thing with Potato.  When it has been running for a
 while (now 11 days but that happened before as well) and I try to
 launch an X client, I got the error message:
 
   _X11TransSocketUNIXConnect: Can't connect: errno = 111
   Gtk-WARNING **: cannot open display: :0.0
 
 The crazy thing is that it is enough to issue
 
   export DISPLAY=localhost:0.0
 
 to get things working again from a shell (DISPLAY used to be only
 :0.0).

Try these commands:

ls -l /tmp/.X11-unix
xhost
xauth list

and post the results if you need help with their output.
The latter two need access to the X server, so you probably should apply
your workaround first.

Regards,
Mirko

Re: _X11TransSocketUNIXConnect: Can't connect: errno = 111

[Mirko Parthey]

On Wed, 16 Aug 2000, Christophe TROESTLER wrote:

 Well, in the meantime I restarted the X server -- needed to do some
 real work[tm]!  I send nonetheless the info -- if you can spot
 something wrong...
 
 $ xauth list
 aiglenoir/unix:0  MIT-MAGIC-COOKIE-1  17afba42b84d5952fce82446e7c3698e
 aiglenoir/unix:0  XDM-AUTHORIZATION-1  ac717264e0f5775dbd46e0d077fc
 jedi.bxlug.be/unix:0  MIT-MAGIC-COOKIE-1  7d266805540e68476949767906173b49
 aiglenoir.linkline.be/unix:0  MIT-MAGIC-COOKIE-1  
 0f16461a07352472131746616c611009
 212.35.0.45:0  MIT-MAGIC-COOKIE-1  4f12270b6423265d2d7e0946651b1a57
 localhost.localdomain/unix:0  MIT-MAGIC-COOKIE-1  
 b8c32f12b2647caeba036f823fcfe0c6
 aiglenoir:0  MIT-MAGIC-COOKIE-1  c65f970f57b10a2767d8a0ef360f31b2

Sorry, I should have told you to censor the keys, because they
are sensitive info. :-(  For your safety, I strongly recommend restarting
the X server on *all* machines mentioned in your listing (if you haven't
done yet after xauth list).

I'm still using slink, but I had a similar problem with some unofficial
X packages. I wasn't able to figure out what the problem was, so I
reverted to official slink X packages. I suppose it had something to do
with confusion about plaintext (MIT-MAGIC-COOKIE-1) vs.
DES (XDM-AUTHORIZATION-1) authorization.

Are you using xdm? Then you might also want to check
  xauth -f /var/lib/xdm/authdir/authfiles/YOUR_FILENAME list
and compare the key(s) to your xauth list output. You can find out the
filename this way:
  ps auxw | grep X.*-auth
it's after the -auth command line option to your X server.

If you need to post keys again, please cut them to 3..4 digits or send
them through md5sum:
  echo key | md5sum
but do so consistently.
The important diagnostic info is whether keys are the same, not their
actual value. 

Otherwise, I'm afraid I can't be of much help.
Maybe someone with more eXperience could take over.

Mirko