Re: partitioning hard drive & /usr is already 96% full

[Urs Thuermann]

Andy <[EMAIL PROTECTED]> writes:

> The only thing I have done is install KDE 3.1 and OpenOffice.org and now
> /usr is 96% full.  (/home is large due to temp. storage from another install)
> 
> Question for the list:
> What is the lists advice in managing my /usr partition
> so it does not completetly fill up and cause problems in the future?

My advice is to use LVM (Logical Volume Manager).  With LVM you don't
use disk partitions for your file systems, but instead create logical
volumes (LV), which can be easily created, deleted and resized
(extended and reduced), without moving data around.  This means you
can easily adapt your file systems sizes to your needs without even
rebooting.

For example, on my machine (not debian though), it looks like this:

isnogud:urs$ df
Filesystem   1k-blocks  Used Available Use% Mounted on
/dev/vg0/root   126931 50529 69849  42% /
/dev/sda1 7988  2213  5775  28% /boot
/dev/vg0/var507748396058 85476  83% /var
/dev/vg0/news  2014611   1620327394284  81% /var/spool/news
/dev/vg0/usr   2579707   1874626574009  77% /usr
/dev/vg0/opt253871 50038193348  21% /opt
/dev/vg0/home  8254992   5201281   2718167  66% /home
/dev/vg0/galois  63461 15471 44714  26% /tftpboot/galois
tmpfs   12892440128884   1% /tmp
/dev/vg0/old 63461 54190  6650  90% /OLD
/dev/vg0/local  515940244416250553  50% /usr/local
/dev/vg0/ftp   1031880989616 42264  96% /usr/local/ftp

When I installed debian on another machine, however, there was AFAIR
no option to use LVM.  So you would have to do this by hand after
installation.


urs


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

Re: partitioning hard drive & /usr is already 96% full

[Urs Thuermann]

Hans Wilmer <[EMAIL PROTECTED]> writes:

> How what about LVM? Can it make things easier? If you can, for
> example, just throw in a new disk, create a partition on it and
> somehow add the additional space to an existing partition, dealing
> with running out of disk space would be easy.

Yes, that is possible with LVM, too.  Just add the new disk, make it a
physical volume and add that physical volume to your volume group.
This would look like this, assuming you add a second SCSI disk,
/dev/sdb.

# fdisk /dev/sdb
/* create one partition /dev/sdb1 covering the whoqle disk */
# pvcreate /dev/sdb1
# vgextend vg0 /dev/sdb1

In fact, it's not even necessary to partition your drive at all,
i.e. you can pvcreate /dev/sdb.


urs


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

Re: crontab on UTC?

[Urs Thuermann]

Nori Heikkinen <[EMAIL PROTECTED]> writes:

> however, every night at 8 PM, the job executes, and not at 2 AM.  i'm 6
> hours behind UTC -- could my crontab be on UTC somehow?  both 'date' and
> 'hwclock' produce the local time output, and i've never seen this
> before.

Is /etc/localtime a symlink to the correct file?  Or is it a copy of a
zoneinfo file?  What does

$ ls -l /etc/localtime
$ ls -lL /etc/localtime

show?  Also, what is the output of

$ date
$ date -u

on your system?


urs


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

Re: 2.2 kernel vs 2.4 kernel

[Urs Thuermann]

Sam Rosenfeld <[EMAIL PROTECTED]> writes:

> What are the advantages (and disadvantages ) of Linux 2.4 compared
> with Linux 2.2?  References will suffice.

* netfilter/iptables instead of ipchains: Better packet filtering,
  e.g. stateful filtering, many more matches on packets with lots of
  kernel modules which can be easily be written.

* Logical Volume Manager (LVM)

* File systems: ext3, reiserfs, tmpfs (RAM based FS that shares the
  swap space), devfs (device file system)

* Bett^WNew VM (virtual memory) system

* Better USB support

I probably missed some points.


urs


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

how to change hostname

[Urs Thuermann]

What is the debain way to change to hostname of a system.  Only
editing /etc/hostname is not sufficient, because there are other
places where the hostname is stored, e.g. /etc/hosts, /etc/mailname,
etc.

Should it be done by running base-config?

urs


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

Re: how to change hostname

[Urs Thuermann]

Maurits van Rees <[EMAIL PROTECTED]> writes:

> I think you just issue the following command as root:
> 
> hostname 

No.  The hostname command only sets the current hostname in the
running kernel.  I want to change the hostname permanently.

urs


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

Re: how to change hostname

[Urs Thuermann]

Lee Braiden <[EMAIL PROTECTED]> writes:

> I *think* the hostname command will set /etc/hostname for you, as well as 
> forcing the running system to update its own idea of what the hostname is.

It only sets the hostname in the kernel using the hostname(2) syscall.

> For /etc/mailname, dpkg-reconfigure exim4-config should do it.
> 
> As for /etc/hosts, I think your options are either to edit it manually, or 
> use 
> some dynamic system like DNS.  However, base-config *might* help.

OK, I will try if base-config does everything I expect it to do.

urs


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

Re: mtab very long

[Urs Thuermann]

Sebastian Witt <[EMAIL PROTECTED]> writes:

> my problem is, that the output of "mount" is very long (and also the
> /etc/mtab file.)
> it has double entries over and over:

Haven't tried it, but I think the following would help:

: > /etc/mtab
mount -a -f


urs


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

Re: want to change /var/log partition to a directory in /var --help

[Urs Thuermann]

Ionut Georgescu <[EMAIL PROTECTED]> writes:

> Please send fdisk -l /dev/hda. Judging by your /etc/fstab, you can't do
> anything because you have /home and /tmp between /var and /var/log. The
> best you cand do is to move /tmp and /var/log to /var and extend /home
> :-)
> 
> I will describe the next steps anyway:
> 
> First: ext3 must resized offline. Second, if you modify the partition
> table while running you have to reboot anyway, because the kernel
> cannot reload the partition table while the disk is being used. You can
> combine these two into 2 steps:
> 
> cp -r /var/log /home/log.bak
> adjust partition table. delete the /var/log line in /etc/fstab
> 
> reboot into single user mode
> 
> What happens: after reboot, /dev/hda3 will be larger, but the filesystem
> has not changed yet. Being in single user mode, no program is running
> at all so you can just
>   umount /var
>   ext2resize /dev/hda3
>   mount /var
> done
> 
> now you just need to copy log
> mount -o remount,rw /home
> rmdir /var/log
> mv /home/log.bak /var/log
> reboot

Alternatively, reinstall the whole system using Logical Volume
Management (LVM), and never fiddle around with partition resizing,
partition moving/copying, not even with partitions at all.  Never again
have to reboot for resizing your file systems, swap spaces or whatever
block devices.

Just call lvextend/lvreduce and resize2fs (or equivalent for your FS)
after umounting the affected file systems and your done.


urs


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

apache2 install problems

[Urs Thuermann]

I tried to install apache2 on a debian testing system.  However, I get
the following error:

# apt-get install apache2
Reading Package Lists... Done
Building Dependency Tree... Done
The following extra packages will be installed:
  apache2-common apache2-mpm-worker
Suggested packages:
  apache2-doc lynx www-browser
The following NEW packages will be installed:
  apache2 apache2-common apache2-mpm-worker
0 upgraded, 3 newly installed, 0 to remove and 1 not upgraded.
Need to get 0B/1084kB of archives.
After unpacking 2613kB of additional disk space will be used.
Do you want to continue? [Y/n] y
Selecting previously deselected package apache2-common.
(Reading database ... 17566 files and directories currently installed.)
Unpacking apache2-common (from .../apache2-common_2.0.51-2_i386.deb) ...
Selecting previously deselected package apache2-mpm-worker.
Unpacking apache2-mpm-worker (from .../apache2-mpm-worker_2.0.51-2_i386.deb) ...
Selecting previously deselected package apache2.
Unpacking apache2 (from .../apache2_2.0.51-2_i386.deb) ...
Setting up apache2-common (2.0.51-2) ...
Setting Apache2 to Listen on port 80. If this is not desired, please edit 
/etc/apache2/ports.conf as desired. Note that the Port directive no longer works.
dpkg: error processing apache2-common (--configure):
 subprocess post-installation script returned error exit status 1
dpkg: dependency problems prevent configuration of apache2-mpm-worker:
 apache2-mpm-worker depends on apache2-common (= 2.0.51-2); however:
  Package apache2-common is not configured yet.
dpkg: error processing apache2-mpm-worker (--configure):
 dependency problems - leaving unconfigured
dpkg: dependency problems prevent configuration of apache2:
 apache2 depends on apache2-mpm-worker (= 2.0.51-2) | apache2-mpm-prefork (= 
2.0.51-2) | apache2-mpm-perchild (= 2.0.51-2);
however:
  Package apache2-mpm-worker is not configured yet.
  Package apache2-mpm-prefork is not installed.
  Package apache2-mpm-perchild is not installed.
dpkg: error processing apache2 (--configure):
 dependency problems - leaving unconfigured
Errors were encountered while processing:
 apache2-common
 apache2-mpm-worker
 apache2
E: Sub-process /usr/bin/dpkg returned an error code (1)


How can I fix this problem?  Or are there other, maybe older apache2
packages that install cleanly?


urs


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

Re: apache2 install problems

[Urs Thuermann]

Marcos Carneiro da Rocha <[EMAIL PROTECTED]> writes:

> Try to apt-get install apache2-mpm-prefork instead of apache2.
> apache2 is a virtual package to another package but there is a conflict
> between apache2-mpm-prefork and apace2-mpm-worker and
> apache2-mpm-perchild - you should choose one of these.

Ah, thanks, that seems to work.


urs


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

Single broken package blocks whole package management

[Urs Thuermann]

After an dist-upgrade from Raspian 8 (jessie) to 9.13 (stretch)
hundreds of packages still need to be upgraded and aptitude reports
numerous conflicts.

I first wanted to upgrade everything which doesn't causes any
conflicts, which fails because of problems in wolfram-engine:

root@uranus:~# aptitude safe-upgrade --no-new-installs -RZ
Resolving dependencies...
open: 2242; closed: 1906; defer: 121; conflict: 311
The following packages will be upgraded:
  acl <+36.9 kB>  adduser <-222 kB>  alacarte <-20.5 kB>  
  alsa-utils <+198 kB>  aspell-en <+95.2 kB>  avahi-daemon <+53.2 kB>  
  [...]
  xserver-common <+53.2 kB>  xserver-xorg <+1024 B>  
  xserver-xorg-input-all <+1024 B>  xterm <+163 kB>  
The following partially installed packages will be configured:
  debconf-utils  
The following packages are RECOMMENDED but will NOT be installed:
  at-spi2-core  bsd-mailx  cinnamon  dunst  exfat-utils  fonts-piboto  
  gdisk  gnome-accessibility-themes  gnome-flashback  gnome-screenshot
  [...]
  xfce4-notifyd  xfonts-base  xserver-xorg-input-wacom  
  xserver-xorg-legacy  
663 packages upgraded, 0 newly installed, 0 to remove and 344 not upgraded.
Need to get 0 B/283 MB of archives. After unpacking 89.8 MB will be used.
Do you want to continue? [Y/n/?] 
Reading changelogs... Done   
Unescaped left brace in regex is deprecated, passed through in regex; 
marked by <-- HERE in m/^(.*?)(\\)?\${ <-- HERE ([^{}]+)}(.*)$/ at 
/usr/share/perl5/Debconf/Question.pm line 72.
Unescaped left brace in regex is deprecated, passed through in regex; 
marked by <-- HERE in m/\${ <-- HERE ([^}]+)}/ at 
/usr/share/perl5/Debconf/Config.pm line 30.
Extracting templates from packages: 100%
Preconfiguring packages ...
dpkg: unrecoverable fatal error, aborting:
 files list file for package 'wolfram-engine' contains empty filename
E: Sub-process /usr/bin/dpkg returned an error code (2)
Failed to perform requested operation on package.  Trying to recover:

root@uranus:~#

I tried to remove packages wolfram-engine and wolframscript, also
tried to remove debconf-utils, but everything fails with the same
error message:

root@uranus:~# dpkg --force-all -P wolfram-engine 
dpkg: unrecoverable fatal error, aborting:
 files list file for package 'wolfram-engine' contains empty filename

Also, upgrading a single package that's completely unrelated, is not
possible:

root@uranus:~# aptitude install acl
The following packages will be upgraded: 
  acl libacl1 
The following partially installed packages will be configured:
  debconf-utils{b} 
2 packages upgraded, 0 newly installed, 0 to remove and 1005 not upgraded.
Need to get 0 B/80.7 kB of archives. After unpacking 49.2 kB will be used.
The following packages have unmet dependencies:
 debconf-utils : Depends: debconf (= 1.5.61) but 1.5.56+deb8u1 is installed 
and it is kept back.
The following actions will resolve these dependencies:

 Remove the following packages:
1) debconf-utils   



Accept this solution? [Y/n/q/?] 
The following packages will be REMOVED:
  debconf-utils{a} 
The following packages will be upgraded:
  acl libacl1 
2 packages upgraded, 0 newly installed, 1 to remove and 1005 not upgraded.
Need to get 0 B/80.7 kB of archives. After unpacking 58.4 kB will be freed.
Do you want to continue? [Y/n/?] 
Reading changelogs... Done   
Unescaped left brace in regex is deprecated, passed through in regex; 
marked by <-- HERE in m/^(.*?)(\\)?\${ <-- HERE ([^{}]+)}(.*)$/ at 
/usr/share/perl5/Debconf/Question.pm line 72.
Unescaped left brace in regex is deprecated, passed through in regex; 
marked by <-- HERE in m/\${ <-- HERE ([^}]+)}/ at 
/usr/share/perl5/Debconf/Config.pm line 30.
dpkg: unrecoverable fatal error, aborting:
 files list file for package 'wolfram-engine' contains empty filename
E: Sub-process /usr/bin/dpkg returned an error code (2)
Failed to perform requested operation on package.  Trying to recover:

root@uranus:~#

What else can I do to get the package management working again?

urs

Re: Single broken package blocks whole package management

[Urs Thuermann]

Richard Hector  writes:

> Firstly, the standard response is that Raspbian is not Debian :-)
> There are differences which might be related to your problem.

Yes, of course.  I know.  But my question was not about the Raspbian
specific packages but on apt, dpkg, aptitude & co.  I assume, these
work mostly (or exactly) like in Debian.

> > I first wanted to upgrade everything which doesn't causes any
> > conflicts, which fails because of problems in wolfram-engine:
> 
> wolfram-engine appears not to be a debian package, for starters.
> 
> [...]
> 
> Ah. So perhaps a bug in that package. Or it's corrupted.
> 
> If you look at /var/lib/dpkg/info/wolfram-engine.list does it have an
> empty line in it? What happens if you edit that out?

OK, thanks for that hint.  This file indeed not only contained a blank
line, but it seems to be completely corrupted.  Roughly 5-10% of all
bytes in random positions are replaced with random (some ASCII, some
non-ASCII) values, making *lots* corrupted path names, which I can
still recognize/guess, but corrupted.  And one line had its last
character replaced by a new-line, thus making a blank line.  A quick
check

# cd /var/lib/dpkg/info; file *.list | grep -v ASCII
ca-certificates.list:   UTF-8 Unicode text
libmodule-build-perl.list:  empty
python-apt-common.list: data
wolfram-engine.list:data

showed another corrupted file.  python-apt-common.list has the same
random errors.  Just curious, I wanted to know, whether all other
files contain valid and existing path names:

# for f in *list; do echo $f; cat $f | tr '\n' '\0' | 
  xargs -0 ls -d >/dev/null ; done 2>&1 | less

Many package list files contain path names which seem valid but are
non-existent in the file system.  Is that normal?  On another
Debian stretch/amd64 system with 1868 packages installed I got only 3
packages each containing only 1 non-existent path name.

> I'd try deleting any blank lines from that file, and trying again.

Yes, I will try that.  Just to learn if it will fix things.  I expect
many errors because of the many broken path names which cannot be
removed.

However, since this system looks somewhat suspicious already and I
wanted to go to buster (or maybe bullseye) anyway, I will save /home
and relevant stuff from /var and then reinstall from scratch.  After
testing the SD card.  Seems to be easier.

> Or maybe apt-get install --reinstall wolfram-engine

It all leads to the same error message as above.
I don't need Wolfram anyway, it's been installed by default.

> Or ask on a Raspbian list :-)

Hm, probably I wasn't clear enough.  My question was not about
Raspbian or the wolfram package, but about dpkg and friends.  Isn't
there a better way to remove packages if the .list file corrupted?
And isn't there a way to temporarily ignore such errors and install or
upgrade other packages despite such an error?  Even dpkg --force-all
didn't help so I couldn't do aynthing.  I found it surpring that a
single broken package leaves you in a state where you're stuck and
cannot go out, easily.

> https://unix.stackexchange.com/questions/425355/x11-common-contains-empty-filename

Looks very similar to my problem.

urs

Re: Single broken package blocks whole package management

[Urs Thuermann]

"Andrew M.A. Cater"  writes:

> Raspbian _isn't_ Debian. Wolfram-engine is a third party (commercial) app -
> Wolfram Mathematica which the Raspberry Pi foundation licences with a 
> special educational arrangement.
> 
> You might be able to force a reinstall of wolfram-engine to produce
> something consistent to then remove but this is probably wasted
> effort.

Yes, and my question should be about dpkg & friends how to get out of
that state where a single corrupted package blocks everything.

> Rather than trying to update from 8 to 9, unless you have significant
> data invested, I would suggest simply downloading a copy of the latest
> release of Raspberry Pi OS based on Debian 11.
> 
> Although I can guarantee that Debian should be upgradable between major
> releases, I can't be sure for other Debian derivatives.

I hoped for the chain of upgrades to work since that would be easier
than installing from scratch.  Since that Raspberry Pi is head-less in
the room next-door.  I need to pull the SD card, write the image onto
it, put the Raspi to some screen at my desk, find a keyboard etc.  Not
much of a problem, but dist-upgrading through remote-login would be
less effort.  Now, I'll do a installation from scratch.

> As ever, for Debian: https://wiki.debian.org/DontBreakDebian

Yes, sure.  AFAIK, nothing of that sort has been done to that Raspi,
at least, as long as I was the only root on it.  I gave to my son a
year ago or so, but I think he didn't anything wrong to it.

> Debian installers which will install something closer to vanilla Debian
> are available - but don't include Wolfram's software. They're unofficial
> in that they're dependnet on non-free software to install. See 
> https://cdimage.debian.org/cdimage/unofficial/raspi/

I think I'll give it a try.  On one of my old Raspis idling in my home
network or lying around and gathering dust.  Especially the 64 bit
Debian should be interesting on my Raspberry Pi 3.  So thank for that
pointer and suggestion.

urs

Re: I've been caught out

[Urs Thuermann]

Greg Wooledge  writes:

> On Mon, Jan 24, 2022 at 05:01:21AM -0500, gene heskett wrote:
> > gene@coyote:~/Debian-arm/linux$ patch -p1  ../patches/*.patch
> 
> That's not how you do it.  patch(1) can only accept one patch at a time,
> and it expects to see it on standard input.
> 
> for p in ../patches/*.patch; do patch -p1 < "$p"; done

You can even do the somewhat easier

cat ../patches/*.patch | patch -p1

which will do the same.  

However, often the order of patches is important when some patch
depends on another being applied before.  Then *.patch will probably
not work since it applies patches in alphabetical order.  That's, why
I also like to work with quilt a lot.  With quilt, the patches/series
file describes the patch series, i.e. the order in which patches have
to be applied.  With "quilt push" and "quilt pop" you can apply and
remove single patches, or go to a specific patch with "quilt push
" (or "quilt pop "), or apply/remove all patches with
"quilt push -a" and "quilt pop -a".  Always in the correct order.

urs

RAID-1 and disk I/O

[Urs Thuermann]

On my server running Debian stretch, the storage setup is as follows:
Two identical SATA disks with 1 partition on each drive spanning the
whole drive, i.e. /dev/sda1 and /dev/sdb1.  Then, /dev/sda1 and
/dev/sdb1 form a RAID-1 /dev/md0 with LVM on top of it.

The disk I/O shows very different usage of the two SATA disks:

# iostat | grep -E '^[amDL ]|^sd[ab]'
Linux 5.13.1 (bit)  07/17/21_x86_64_(2 CPU)
avg-cpu:  %user   %nice %system %iowait  %steal   %idle
   3.780.002.270.860.00   93.10
Device:tpskB_read/skB_wrtn/skB_readkB_wrtn
sdb   4.5472.1661.25   54869901   46577068
sda   3.7235.5361.25   27014254   46577068
md0   5.53   107.1957.37   81504323   43624519

The data written to the SATA disks is about 7% = (47 GB - 44 GB) / 44 GB
more than to the RAID device /dev/md0.  Is that the expected overhead
for RAID-1 meta data?

But much more noticable is the difference of data reads of the two
disks, i.e. 55 GB and 27 GB, i.e. roughly twice as much data is read
from /dev/sdb compared to /dev/sda.  Trying to figure out the reason
for this, dmesg didn't give me anything but I found the following with
smartctl:

--
# diff -U20 <(smartctl -x /dev/sda) <(smartctl -x /dev/sdb)
--- /dev/fd/63  2021-07-17 12:09:00.425352672 +0200
+++ /dev/fd/62  2021-07-17 12:09:00.425352672 +0200
@@ -1,165 +1,164 @@
 smartctl 6.6 2016-05-31 r4324 [x86_64-linux-5.13.1] (local build)
 Copyright (C) 2002-16, Bruce Allen, Christian Franke, www.smartmontools.org
 
 === START OF INFORMATION SECTION ===
 Model Family: Seagate Barracuda 7200.14 (AF)
 Device Model: ST2000DM001-1ER164
-Serial Number:W4Z171HL
-LU WWN Device Id: 5 000c50 07d3ebd67
+Serial Number:Z4Z2M4T1
+LU WWN Device Id: 5 000c50 07b21e7db
 Firmware Version: CC25
 User Capacity:2,000,397,852,160 bytes [2.00 TB]
 Sector Sizes: 512 bytes logical, 4096 bytes physical
 Rotation Rate:7200 rpm
 Form Factor:  3.5 inches
 Device is:In smartctl database [for details use: -P show]
 ATA Version is:   ACS-2, ACS-3 T13/2161-D revision 3b
 SATA Version is:  SATA 3.1, 6.0 Gb/s (current: 3.0 Gb/s)
 Local Time is:Sat Jul 17 12:09:00 2021 CEST
 SMART support is: Available - device has SMART capability.
 SMART support is: Enabled
 AAM feature is:   Unavailable
 APM level is: 254 (maximum performance)
 Rd look-ahead is: Enabled
 Write cache is:   Enabled
 ATA Security is:  Disabled, NOT FROZEN [SEC1]
 Wt Cache Reorder: Unavailable
 
 === START OF READ SMART DATA SECTION ===
 SMART overall-health self-assessment test result: PASSED
 
 General SMART Values:
 Offline data collection status:  (0x82)Offline data collection activity
was completed without error.
Auto Offline Data Collection: Enabled.
 Self-test execution status:  (   0)The previous self-test routine 
completed
without error or no self-test has ever 
been run.
 Total time to complete Offline 
-data collection:   (   89) seconds.
+data collection:   (   80) seconds.
 Offline data collection
 capabilities:   (0x7b) SMART execute Offline immediate.
Auto Offline data collection on/off 
support.
Suspend Offline collection upon new
command.
Offline surface scan supported.
Self-test supported.
Conveyance Self-test supported.
Selective Self-test supported.
 SMART capabilities:(0x0003)Saves SMART data before entering
power-saving mode.
Supports SMART auto save timer.
 Error logging capability:(0x01)Error logging supported.
General Purpose Logging supported.
 Short self-test routine 
 recommended polling time:   (   1) minutes.
 Extended self-test routine
-recommended polling time:   ( 213) minutes.
+recommended polling time:   ( 211) minutes.
 Conveyance self-test routine
 recommended polling time:   (   2) minutes.
 SCT capabilities: (0x1085) SCT Status supported.
 
 SMART Attributes Data Structure revision number: 10
 Vendor Specific SMART Attributes with Thresholds:
 ID# ATTRIBUTE_NAME  FLAGSVALUE WORST THRESH FAIL RAW_VALUE
-  1 Raw_Read_Error_Rate POSR--   119   099   006-208245592
-  3 Spin_Up_TimePO   097   096   000- 

Re: RAID-1 and disk I/O

[Urs Thuermann]

David Christensen  writes:

> You should consider upgrading to Debian 10 -- more people run that and
> you will get better support.

It's on my TODO list.  As well as upgrading the very old hardware.
Currently, it's a Gigabyte P35-DS3L with an Intel Core2Duo E8400 CPU
and 8 GB RAM.  It's only my private home server and performance is
still sufficient but I hope to reduce power consumption considerably.

> > the storage setup is as follows:
> > Two identical SATA disks with 1 partition on each drive spanning the
> > whole drive, i.e. /dev/sda1 and /dev/sdb1.  Then, /dev/sda1 and
> > /dev/sdb1 form a RAID-1 /dev/md0 with LVM on top of it.
> 
> 
> ext4?  That lacks integrity checking.
> 
> 
> btrfs?  That has integrity checking, but requires periodic balancing.

Mostly ext4 for / /var /var/spool/news /usr /usr/local and /home file
systems.  The /usr/src file system is btrfs and some test file systems
also.  There are also 4 VMs, FreeBSD and NetBSD with their partitions
and slices and ufs file systems, one Linux VM with ext4 and one very
old Linux VM (kernel 2.4) with its own LVM in two LVs and 10 ext3 file
systems.

> Are both your operating system and your data on this array?  I always
> use a single, small solid-state device for the system drive, configure
> my hardware so that it is /dev/sda, and use separate drive(s) for data
> (/dev/sdb, /dev/sdc, etc.).  Separating these concerns simplifies
> system administration and disaster preparedness/ recovery.

Yes, everything is in the LVs on /dev/md0.  Except for some external
USB hard drives for backup (4 TB) and some other seldomly used stuff
(e.g. NTFS drive with some old data of my wife's laptop, I cannot
persuade her to use Linux).

> > but I found the following with
> > smartctl:
> > --
> > # diff -U20 <(smartctl -x /dev/sda) <(smartctl -x /dev/sdb)
> 
> 
> Why limit unified context to 20 lines?  You may be missing information
> (I have not counted the differences, below).  I suggest '-U' alone.

20 lines are just enough to get all.  You can see this because there
are less than 20 context lines at the beginning and end of the diff
and only one hunk.  GNU diff doesn't allow -U without a line count.

> You have a SATA transfer speed mismatch -- 6.0 Gbps drives running at
> 3.0 Gbps.  If your ports are 3 Gbps, fine.  If your ports are 6 Gbps,
> you have bad ports, cables, racks, docks, trays, etc..

Yes, the old Gigabyte mainboard has only 3 Gbps ports.  I wasn't aware
of this but have just looked up the specs.

> Seek_Error_Rate indicates those drives have seen better days, but are
> doing their job.
> 
> 
> Power_On_Hours indicates those drives have seen lots of use.

> Power_Cycle_Count indicates that the machine runs 24x7 for long
> periods without rebooting.

Yes, the server runs 24/7 except for kernel updates, and a power
outage 2 weeks ago (my UPS batteries also need replacement... )-:

And the server is about 8 years old, initially with only 1 hard drive
which crashed while my backup was too small to hold everything.  This
meant a lot of work (and quite some money) to get everything running
again and to recover data which wasn't in the backup.

This was almost 6 years ago and I then bought 2 Seagate Barracuda
drives for RAID-1 and a larger backup drive.  One of the two Seagate
drives is still running and is /dev/sda.  The other drive /dev/sdb
crashed after only 9.5 months of operation and I got it replaced by
the dealer.  This was when I loved my decision to setup RAID-1.  With
no downtime I pulled the failed drive, returned it to the dealer, ran
the system a week or two with only one drive, got the replacement
drive from the dealer hot-plugged it in, synced, and was happy :-)
Only short time after this I also bought a 3.5" removable mounting
frame for 2 drives to swap drives even more easily.

> Runtime_Bad_Block looks acceptable.

> End-to-End_Error and Reported_Uncorrect look perfect.  The drives
> should not have corrupted or lost any data (other hardware and/or
> events may have).

OK.

> Airflow_Temperature_Cel and Temperature_Celsius are higher than I
> like. I suggest that you dress cables, add fans, etc., to improve
> cooling.

OK, I'll have a look at that.

> UDMA_CRC_Error_Count for /dev/sda looks worrisome, both compared to
> /dev/sdb and compared to reports for my drives.
> 
> 
> Total_LBAs_Written for /dev/sda is almost double that of
> /dev/sdb. Where those drives both new when put into RAID1?

Yes, see above.  But /dev/sdb was replaced after 9.5 months, so it has
shorter life-time.  Also, /dev/sda began to fail every couple of
months about a year ago.  I could always fix this by pulling the
drive, re-inserting and re-syncing it.  This also caused more
write-traffic to /dev/sda.

> > SMART Extended Self-test Log Version: 1 (1 sectors)
> >   Num  Test_DescriptionStatus  Remaining  
> > LifeTime(hours)  LBA_of_first_error
> > -# 1  Short o

After uprade to buster: QEMU doesn't support -drive if=scsi

[Urs Thuermann]

After uprading from stretch to buster on my server (x86_64), I cannot
start one of my VMs anymore.  The guest is a *very* old Linux (non
Debian, i686, kernel 2.4, GRUB 1.99) and uses the Symbios Logic
sym53c8xx_2 SCSI driver in its init-ramdisk to mount the root file
system and another virtual disk containing LVM.  The init-ramdisk also
has an IDE driver.

The newer QEMU seems not to support this virtual SCSI adaptor:

# kvm -cpu pentium3 ... \
  -drive 
file=/dev/vg0/.sda,format=raw,if=scsi,media=disk,cache=writeback" \
  -drive 
file=/dev/vg0/.sdb,format=raw,if=scsi,media=disk,cache=writeback" \
qemu-system-x86_64: -drive 
file=/dev/vg0/.sda,format=raw,if=scsi,media=disk,cache=writeback: 
machine type does not support if=scsi,bus=0,unit=0
qemu-system-x86_64: -drive 
file=/dev/vg0/.sdb,format=raw,if=scsi,media=disk,cache=writeback: 
machine type does not support if=scsi,bus=0,unit=1

To be able to start the VM, I have changed the options -drive ... to
-hda and -hdb to use the IDE drivers in the guest.  This works
(somewhat) but performance is *extremely* poor.  Several 10 times
slower than with SCSI.

Another problem, which I don't know whether it's related to the disk
I/O performance, is that ntpd is not even able to sync the system
clock.  Even when there is almost no disk I/O.

Is there any chance of getting this guest running with the old SCSI
adaptor in this new version of QEMU in buster?

I could also try to compile a new Linux 2.4 kernel and build a new
init-ramdisk.  But which driver should I use? In Linux 2.4 there is no
virtio.

In the long run I should clearly move the whole thing to a newer
Debian Linux machine.  But as I am short of time currently, I'd like
to be able to continue with this old VM, at least for a while.

Steve

texlive-binaries depends on 2 versions of a library

[Urs Thuermann]

To me this looks somewhat surprising that in Debian buster a package
depends on 2 versions of a library:

# apt-cache depends texlive-binaries |grep lua
  Depends: libtexlua52
  Depends: libtexlua52
  Depends: libtexlua53
  Depends: libtexlua53
  Depends: libtexluajit2
  Depends: libtexluajit2
  Breaks: 
  Replaces: 

Is that really intended that it depends on both, libtexlua52 and
libtexlua53?

Steve

Re: /home as a symlink?

[Urs Thuermann]

Jesper Dybdal  writes:

> Thanks for your response.  That would be the natural way of doing it
> if I were partitioning a new disk.  But I don't want to do that, and
> the target disk also has other data, so /home cannot be a complete
> partition.

IIUC, you have a directory on that disk where you want to move the
home directories of your users i.e. /some/path/to/homes to, as well as
some some other directories on that disk.

You could then mount that disk to some mount point, say /data, and
then mount --bind /data/some/path/to/homes /home.

urs

Re: Don't disable recoomends by default

[Urs Thuermann]

Jonas Smedegaard  writes:

> Quoting Stephan Seitz (2019-07-12 09:30:38)
> > On Fr, Jul 12, 2019 at 09:13:29 -0300, Jonas Smedegaard wrote:
> > >Wrong.  Suggests are for packages useful only "sometimes", recommends
> > >are for pacakges needed in "all but unusual installations."
> > 
> > From my experience this is wrong.
> > 
> > With recommends my d10 update would have systemd as init instead of 
> > sysvinit. And I would have got (for example) the package debsecan 
> > which I don’t need.
> > 
> > So it is better to disable recommends and look at the recommended 
> > packages.
> 
> There is nothing wrong in suppressing specific recommendations where you 
> understand the implications.
> 
> What is wrong is to suppress all recommendations by default.

I have done this for years now, i.e. I have

APT::Install-Recommends "false";

in /etc/apt/apt.conf and I haven't had any problems, unexpected
behavior or non-functional software packages.

OTOH, IMHO many packages have recommendations I don't like, especially
if "Recommends:" means it is recommended except in unusual
installations.

For example a number of packages (e.g. geeqie) recommend cups or lpr
although it's not that unusual to not have a printer.

For some reason I don't know, nfs-common recommends python.  I have
used NFS for >25 years now, most of the time without Python and
without any problems.  And why does it recommend python (i.e. 2.7)
instead of python3?

Why does nut-client recommend bash-completion?  It's not that unusual
that admins work with ksh or zsh or some other shell only and
therefore don't need bash support.

And auctex recommends a list of PDF viewers (or okular which can show
DVI files) while I know still many users who work with xdvi only.  So
using auctex to write TeX files should not imply the user also wants
PDF or okular.

Some packages recommend specific mail readers (e.g. logrotate or
smartmontools) when they only need a MTA to send the mail.  Where the
mail is forwarded to and which MUA the admin uses to read it should be
out of the scope of those packages.  The package 'at' does it right by
only recommending mail-transport-agent.

Sometimes a package foo recommends foo-doc while others only suggest
foo-doc.

I could easily extend this list with a dozen more examples.
Installing all recommended packages would increase my installation by
a couple 100 MB, maybe GB.  And this is not only wasted disk
space but also increases the probability of security bugs, which is
bad if the specific packages is actually unused.

But even more annoying is that some packages specify Depends: on other
packages where this should not be the case, IMHO.  This is why I have
cups, bluetooth libs, bluray libs, IEEE1394 libs, MIDI libs, braille
display libs, libs for Wacom tablets, and several dozens of unused
packages on my server machine which has no audio, no bluetooth, no DVD
or Bluray drive, no braille terminal, no printer, no firewire, no
graphical display, ... nothing except a 80x25 text console, lots of
disks, and an Ethernet port.

I would like if many of these packages wouldn't link against these
libraries (and therefore depend on it) but would instead only call
dlopen(3) when the functionality in question is actually used (and
recommend or suggest those libs), and give a diagnostic message if the
lib is missing.

urs

Re: Don't disable recoomends by default

[Urs Thuermann]

Jonas Smedegaard  writes:

> Quoting John Crawley (2019-07-12 22:52:55)
> 
> > Anyway, even if your system default is to install Recommends, apt-get 
> > (and apt too?) always gets user approval before installing anything 
> > beyond the package asked for. If the list looks too long (s)he can 
> > always hit (N) and try again with 'apt-get install 
> > --no-install-recommends '
> 
> ...and suppressing recommends exeptionally like that is quite sensible.
> 
> Again, my strong criticism is only suppressing recommends by *default*.

Before I had suppressed recommends by default in apt.conf I *always*
had to hit 'N' and try again with --no-install-recommends, because
everytime "aptitude full-upgrade" (i.e. no new installs) wanted to
install >100 MB, often 300MB+ of packages I don't want.

urs

aptitude doesn't remove unused packages

[Urs Thuermann]

Debian's package management should remove packages that were installed
automatically, if they are no longer needed.  Unfortunately, that
often seems to not work correctly.  See this example on my Raspberry
Pi running Raspbian jessie:

Some time in the past there were probably packages that needed
gcc-4.8-base so it were installed automatically:

# aptitude show gcc-4.8-base
Package: gcc-4.8-base
State: installed
Automatically installed: yes
Multi-Arch: same
Version: 4.8.4-1
Priority: required
Section: libs
Maintainer: Debian GCC Maintainers 
Architecture: armhf
Uncompressed Size: 176 k
Breaks: dehydra (<= 0.9.hg20110609-2), gcc-4.4-base (< 4.4.7), gcc-4.7-base (< 
4.7.3),
gcj-4.4-base (< 4.4.6-9~), gcj-4.6-base (< 4.6.1-4~), gnat-4.4-base (< 
4.4.6-3~),
gnat-4.6 (< 4.6.1-5~)
Description: GCC, the GNU Compiler Collection (base package)
 This package contains files common to all languages and libraries contained in 
the GNU Compiler
 Collection (GCC).
Homepage: http://gcc.gnu.org/

But after many updates/upgrades no installed package needs
gcc-4.8-base anymore but it hasn't been removed.  If I check why it's
still installed I get:

# aptitude why gcc-4.8-base
i   cron  Recommends exim4 | postfix | mail-transport-agent
p   courier-mta   Provides   mail-transport-agent
p   courier-mta   Dependscourier-authlib (>= 0.66.1-1)
p   courier-authlib   Dependsexpect
p   expectRecommends tk8.6
p   tk8.6 Recommends xterm | x-terminal-emulator
p   gnome-terminalProvides   x-terminal-emulator
p   gnome-terminalRecommends gvfs
p   gvfs  Dependsgvfs-daemons (< 1.22.2-1+deb8u1.1~)
p   gvfs-daemons  Recommends policykit-1-gnome
p   lxsession Provides   policykit-1-gnome
p   lxsession Recommends openbox | x-window-manager
p   xmonadProvides   x-window-manager
p   xmonadRecommends libghc-xmonad-dev
p   libghc-xmonad-dev Dependslibghc-unix-dev-2.6.0.1-7550b
p   ghc   Provides   libghc-unix-dev-2.6.0.1-7550b
p   ghc   Dependslibgmp-dev
p   libgmp-devRecommends libstdc++-4.8-dev | libstdc++-dev
p   libstdc++-4.8-dev Dependsgcc-4.8-base (= 4.8.4-1)

That is there is one package (cron) installed that only recommends one
of a list of packages none of which are actually installed.  Then
there is a long list of Provides, Recommends, and Depends, all of
which are pointless since the root of all this isn't installed.  At
the end of that long list is gcc-4.8-base and that makes aptitude
think it should keep that package?  Strange.
(BTW, this is a router running a very minimal installation, therefore
I don't want any X11, graphical, desktop, window-manager, Haskell, C++
devel stuff on it.)

Even calling apt-get autoremove explicitly doesn't remove that
package:

# apt-get autoremove
Reading package lists... Done
Building dependency tree
Reading state information... Done
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

I regularly observe things like this, not only on Raspbian, but also
on Debian stretch for amd64 which runs on my server machine.

urs

Re: aptitude doesn't remove unused packages

[Urs Thuermann]

The Wanderer  writes:

> if you
> want to know "what's the dependency chain which is keeping this from
> being safe to remove?", you're probably better off running 'apt-get
> --dry-run remove [packagename]', and seeing whether the result wants to
> remove anything that you care about.

Well, I usually just call aptitude purge without --simulate because I
want unneeded packages to be really removed and aptitude asks for
confirmation if it would need to remove more packages.  Here, with -s,
but you can see that gcc-4.8-base is really not needed by anything.

# aptitude remove -s gcc-4.8-base
The following packages will be REMOVED:  
  gcc-4.8-base 
0 packages upgraded, 0 newly installed, 1 to remove and 0 not upgraded.
Need to get 0 B of archives. After unpacking 176 kB will be freed.
Would download/install/remove packages.

> 'apt-get autoremove' deals only with packages which are marked as
> auto-installed; for undetermined reasons, some packages which you
> wouldn't expect to be get marked as manual. 'apt-mark showmanual' will
> list all packages which are flagged as manually-installed.

But as I showed in my first mail, gcc-4.8-base *is* auto-installed:

# apt-mark showmanual | grep gcc
#

Despite this, apt-get autoremove doesn't remove it.

> One thing I habitually do nowadays, to minimize this type of problem, is
> to also run
> 
> # apt-get remove $(deborphan)
> 
> and interleave that back and forth with 'apt-get autoremove' calls until
> neither one of them wants to remove anything anymore.

That also doesn't help:

# apt-get autoremove
Reading package lists... Done
Building dependency tree   
Reading state information... Done
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
# deborphan
raspi-copies-and-fills



BTW, this also is strange since raspi-copies-and-fills is *not*
auto-installed:

# apt-mark showmanual | grep raspi
raspi-copies-and-fills


That's why from time to time I look through the whole list of
installed packages that look suspicious, i.e. auto-installed packages,
especially with names matching lib*, *-common, and *-base, and that
have multiple instances installed of the same name differing only in
version.  It seems there is no other reliable way to get rid of some
old stuff :-(

> This doesn't 100% address the issue, but it will deal with more of it
> than you might think.
> 
> Does the 'deborphan' command report gcc-4.8-base, in your case?

No, as shown above.

urs

After software update systemd runs into timeouts

[Urs Thuermann]

I run a server with Debian 9 and haven't resisted the conversion to
systemd strong enough, therefore I now have a couple of problems with
stability and availability.

One of them is, that often after software updates systemd-logind.service
fails to start and ssh logins take a very long time.  I suspect this has
to do with shared libs being replaced while being used by running
processes.  This usually causes no problems for everything else except
systemd.  When that problem occurs, using lsof I find and restart every
service using deleted files.  Sometimes this helps but not always.

When I try to restart systemd-logind.service it fails with an error
message after a timeout of 1:30 minutes, but then doesn't give any
information about the problem (instead it prefers to show off how little
resources it used and cluttering up output with static info like URLs,
both of which don't belong in a status output):

# time systemctl restart systemd-logind.service 
Job for systemd-logind.service failed because a timeout was exceeded.
See "systemctl status systemd-logind.service" and "journalctl -xe" for details.

real1m30.270s
user0m0.003s
sys 0m0.004s
# systemctl status systemd-logind.service 
● systemd-logind.service - Login Service
   Loaded: loaded (/lib/systemd/system/systemd-logind.service; static; vendor 
preset: enabled)
   Active: activating (start) since Wed 2020-08-05 11:06:06 CEST; 51s ago
 Docs: man:systemd-logind.service(8)
   man:logind.conf(5)
   http://www.freedesktop.org/wiki/Software/systemd/logind
   http://www.freedesktop.org/wiki/Software/systemd/multiseat
 Main PID: 29560 (systemd-logind)
   Status: "Processing requests..."
Tasks: 1 (limit: 4915)
   Memory: 1.0M
  CPU: 10ms
   CGroup: /system.slice/systemd-logind.service
   └─29560 /lib/systemd/systemd-logind

Aug 05 11:06:06 bit systemd-logind[29560]: New session 3893 of user urs.
Aug 05 11:06:06 bit systemd-logind[29560]: New session 3889 of user urs.
...


The log messages in /var/log/syslog also don't give enough information
about what has failed, but it seems systemd tries to restart that service
in the background:

Aug  5 11:06:06 bit systemd[1]: systemd-logind.service: Start operation timed 
out. Terminating.
Aug  5 11:06:06 bit systemd[1]: Failed to start Login Service.
Aug  5 11:06:06 bit systemd[1]: systemd-logind.service: Unit entered failed 
state.
Aug  5 11:06:06 bit systemd[1]: systemd-logind.service: Failed with result 
'timeout'.
Aug  5 11:06:06 bit systemd[1]: systemd-logind.service: Service has no hold-off 
time, scheduling restart.
Aug  5 11:06:06 bit systemd[1]: Stopped Login Service.
Aug  5 11:06:06 bit systemd[1]: Starting Login Service...

# grep Start /var/log/syslog | tail -10
Aug  5 11:37:41 bit systemd[1]: systemd-logind.service: Start operation timed 
out. Terminating.
Aug  5 11:37:41 bit systemd[1]: Starting Login Service...
Aug  5 11:39:11 bit systemd[1]: systemd-logind.service: Start operation timed 
out. Terminating.
Aug  5 11:39:11 bit systemd[1]: Starting Login Service...
Aug  5 11:40:41 bit systemd[1]: systemd-logind.service: Start operation timed 
out. Terminating.
Aug  5 11:40:41 bit systemd[1]: Starting Login Service...
Aug  5 11:42:11 bit systemd[1]: systemd-logind.service: Start operation timed 
out. Terminating.
Aug  5 11:42:11 bit systemd[1]: Starting Login Service...
Aug  5 11:43:41 bit systemd[1]: systemd-logind.service: Start operation timed 
out. Terminating.
Aug  5 11:43:41 bit systemd[1]: Starting Login Service...


With jounalctl you also see the bloated logs which make them unpleasant to
search and it doesn't have more info than /var/log/syslog.

I had this problem 5 to 10 times until now, I have tried fixing it
everytime for an hour or so, and finally given up rebooting the machine.

Is there anything one can do to make systemd working properly again?


In my 30 years of experience with Unix I have always loved the simplicity,
elegance, and stability.  With Linux in general and specifically Debian
with apt, I liked how easily you can update software without disturbing
the running system -- until the advent of systemd.  It makes everything
more complicate, non-Unix-ish, and less stable.  In my experience, systemd
is the end to "Debian as the maintainable OS".

Now I have the choice of not updating, rebooting after most updates, or
having to wait a minute or so for each remote login.  All unacceptable for
a server with long-running processes, hosting a couple of virtual
machines, and high uptimes.


urs

Re: After software update systemd runs into timeouts

[Urs Thuermann]

Greg Wooledge  writes:

> Google led me to 
> which, although it's for Arch Linux and not Debian, might possibly
> be relevant for you.
> 
> In particular, what does "ls -ld /var/run" say?

The symlink /var/run -> /run is OK.  My problem is probably the same
as in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=770135 and
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=823987 which is
nearly 6 years old.  Unfortunately, it seems not to be solved.

urs

Re: After software update systemd runs into timeouts

[Urs Thuermann]

Andrei POPESCU  writes:

> The package 'needrestart' hooks into APT and offers to restart any 
> affected services.

Thx, I didn't know needrestart until today, but learned about in in
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=770135 and
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=823987.

But regardless of using needrestart or finding grepping them from the
output of lsof, it seems systemd has problems with it or dbus being
restarted.

The bug is obviously nearly 6 years old but hasn't been fixed yet.

I have spent much time today restarting dbus, ssh, and/or several
systemd services in different order.  Also systemctl daemon-reexec.
Nothing seems to help and it seems systemd bugs again force me to
reboot.  But I am currently 700 km away from my server for the next
1.5 weeks and noone else has access to it.  So I'd prefer not to risk
it.

urs

Re: After software update systemd runs into timeouts

[Urs Thuermann]

Andrei POPESCU  writes:

> It's of course possible to reopen the bug if it still exists, though 
> stretch is oldstable now and in LTS mode.

I plan to upgrade to buster, but only after August 15th.

> On a server it might be possible to simply not use systemd-logind.  E.g.  
> (at least on buster) libpam-systemd is only recommended by systemd and 
> openssh-server.

Although 'aptitude why' only shows libpam-systemd is recommended by
systemd, I cannot purge it since other packages seem to depend on it:

# aptitude why libpam-systemd
i   systemd Recommends libpam-systemd
# aptitude purge libpam-systemd
The following packages will be REMOVED:
  libpam-systemd{p}
0 packages upgraded, 0 newly installed, 1 to remove and 0 not upgraded.
Need to get 0 B of archives. After unpacking 379 kB will be freed.
The following packages have unmet dependencies:
 policykit-1 : Depends: libpam-systemd but it is not going to be installed
The following actions will resolve these dependencies:

 Remove the following packages:
1) libvirt-daemon-system [3.0.0-4+deb9u4 (now, oldstable)]
2) policykit-1 [0.105-18+deb9u1 (now, oldstable)]

 Leave the following dependencies unresolved:
3) needrestart recommends libpam-systemd | sysvinit-core
4) openssh-server recommends libpam-systemd
5) systemd recommends libpam-systemd
6) virt-manager recommends libvirt-daemon-system (>= 1.2.7)
7) xserver-xorg-core recommends libpam-systemd



Accept this solution? [Y/n/q/?]

Should this be considered a bug?  Shouldn't 'aptitude why' show the
packages that depend on it?

urs

aptitude why (Was: Re: After software update systemd runs into timeouts)

[Urs Thuermann]

David Wright  writes:

> On Wed 05 Aug 2020 at 22:53:26 (+0200), Urs Thuermann wrote:
> 
> > Should this be considered a bug?  Shouldn't 'aptitude why' show the
> > packages that depend on it?
> 
> Why not read the man page:
> 
>Note
>aptitude why does not perform full dependency
>resolution; it only displays direct relationships
>between packages. For instance, if A requires B, C
>requires D, and B and C conflict, “aptitude why-not D”
>will not produce the answer “A depends on B, B
>conflicts with C, and D depends on C”.
> 
>By default aptitude outputs only the "most installed,
>strongest, tightest, shortest" dependency chain. That is,
>it looks for a chain that only contains packages which are
>installed or will be installed; it looks for the strongest
>possible dependencies under that restriction; it looks for
>chains that avoid ORed dependencies and Provides; and it
>looks for the shortest dependency chain meeting those
>criteria. These rules are progressively weakened until a
>match is found.

This is exactly what made me think it is a bug.  aptitude should show
the strongest chain, and "Depends" is stronger than "Recommends".  But
aptitude shows a "Recommends" although a "Depends" exists:

$ aptitude why libpam-systemd
i   systemd Recommends libpam-systemd
$ aptitude why policykit-1 libpam-systemd
i A policykit-1 Depends libpam-systemd

But now I see reason: policykit-1 is also installed only because
virt-manager *recommends* libvirt-daemon-system which depends on
policykit-1:

$ aptitude why virt-manager libpam-systemd
i   virt-manager  Recommends libvirt-daemon-system (>= 1.2.7)
i A libvirt-daemon-system Dependspolicykit-1
i A policykit-1   Dependslibpam-systemd

But since that chain is longer, aptitude shows the recommendation of
systemd.  In the output of aptitude purge in my other mail I hadn't
realized that libvirt-daemon-system is also only automatically
installed.

But thanks anyway for pointing out these sections of the man page.  I
wasn't aware of all the details and now understand better what
aptitude does.

urs

Re: Disks renamed after update to 'testing'...?

[Urs Thuermann]

"Rick Thomas"  writes:

> The /dev/sdx names for devices have been unpredictable for quite a
> while.  Which one is sda and which sdb will depend on things like
> timing -- which one gets recognized by the kernel first.
> 
> The best solution is to either use UUID or LABEL when you fsck
> and/or mount the device.

IMO the best solution is to use LVM.  I use it since 2001 on most
drives and I don't have partitions.  And I prefer to use device names
over using the *UUID or *LABEL prefixes.  With LVM, device names are
predictable /dev/mapper/- with symlinks
/dev//.

urs

Re: Disks renamed after update to 'testing'...?

[Urs Thuermann]

David Christensen  writes:

> AIUI the OP was mounting an (external?) drive partition for use as a
> destination for backups.  Prior to upgrading to Testing, the root
> partition was /dev/sda1 (no LVM?) and the backup partition was
> /dev/sdb1 (no LVM?).  After upgrading to Testing, the root partition
> is /dev/sdb1 and the backup partition device node is unknown.  The OP
> was confused by the changed root partition device node.
> 
> 
> Please describe how LVM would help in this situation.

Instead of using /dev/sdb1 directly for the backup file system, the OP
could put LVM to /dev/sdb1 (or now /dev/sda1).  I.e. he would create a
physical volume on /deb/sdb1, create a volume group e.g. named vgbkup,
and would then create a logical volume, e.g. named lv1.  The device
name for the backup file system would then always be /dev/vgbkup/lv1
regardless of how the kernel will name underlying device (/dev/sda1 or
/dev/sdb1 or whatever).

In addition, you get the flexibility of LVM of adding, deleting, and
resizing volumes without re-partitioning.

If the OP had done that before he hadn't noticed the change from
/dev/sdb1 to /dev/sda1 as he hadn't used that name.  He could now
change to LVM and never deal with changing physical device names
again.

Whether it's an internal or external drive doesn't matter.

My backup drive is an external USB-3 hard drive with 1 partition
containing the whole disk space of 4 TB.  That partition contains
volume vgroup "vg2" with currently one logical voulume "snap" of 2 TB
and I mount /dev/vg2/snap to /var/snapshots for backups.  Currently I
don't need more than these 2 TB but I could easily extend the logical
volume or create new ones for backups of my other machines or virtual
machines.

And I actually forget the physical device name of the vg2 volume group
(I have just looked it up using vgdisplay(8), and it's currently
/dev/sdf1).

urs

Re: Disks renamed after update to 'testing'...?

[Urs Thuermann]

Urs Thuermann  writes:

> IMO the best solution is to use LVM.  I use it since 2001 on most
> drives and I don't have partitions.  And I prefer to use device names
> over using the *UUID or *LABEL prefixes.  With LVM, device names are
> predictable /dev/mapper/- with symlinks
> /dev//.

Following up myself: The reason I prefer stable device names instead
of UUIDs or LABELs is that device names show up in some places even if
you use UUID or LABEL in /etc/fstab or in your command line:

On my laptop I have UUID in /etc/fstab but df still shows the device
name:

$ grep -w / /etc/fstab
# / was on /dev/nvme0n1p2 during installation
UUID=c73ff331-0ff5-44fb-8aef-228e64a96175 /   ext4
errors=remount-ro 0   1
$ df | grep -w /
/dev/nvme0n1p2  237470384 107725172 117659352  48% /

On my server with LVM I get:

$ grep -w / /etc/fstab
/dev/mapper/vg0-root   /   ext4errors=remount-ro 0   1
$ df | grep -w /
/dev/mapper/vg0-root  2031440 659132   1261556  35% /


urs

Re: Disks renamed after update to 'testing'...?

[Urs Thuermann]

David Christensen  writes:

> Thanks for the explanation.  It seems that pvcreate(8) places an LVM
> disk label and an LVM metadata area onto disks or partitions when
> creating a PV; including a unique UUID:
> 
> https://www.man7.org/linux/man-pages/man8/pvcreate.8.html

Yes, correct.  You can see the UUID with pvdisplay(8) or blkid(8):

# pvdisplay /dev/md0
  --- Physical volume ---
  PV Name   /dev/md0
  VG Name   vg0
  PV Size   1.82 TiB / not usable 3.00 MiB
  Allocatable   yes 
  PE Size   4.00 MiB
  Total PE  476899
  Free PE   96653
  Allocated PE  380246
  PV UUID   uFHSzs-QpCa-GVIX-LKRZ-rIRV-KgfE-taQXQV
   
# blkid /dev/md0
/dev/md0: UUID="uFHSzs-QpCa-GVIX-LKRZ-rIRV-KgfE-taQXQV" TYPE="LVM2_member"

> When using a drive as backup media, are there likely use-cases that
> benefit from configuring the drive with no partition, a single PV,
> single VG, single LV, and single filesystem vs. configuring the drive
> with a single partition, single UUID fstab entry, and single
> filesystem?

You can use a partition or the whole disk for a physical volume, as
you can for a file system.  That is, you can

mkfs /dev/sdaormkfs /dev/sda1

and likewise with LVM you can

pvcreate /dev/sdaorpvcreare /dev/sda1

Long ago I actually created PVs on the whole disk and didn't have
partition tables and therefore no partition on many of my drives.
Today, I prefer having a partition table with only one partition
covering the whole disk.  The partition table entry includes a type so
that there is less guessing about what the disk contains:

# fdisk -l /dev/sda | grep /dev
Disk /dev/sda: 1.8 TiB, 2000397852160 bytes, 3907027055 sectors
/dev/sda12048 3907026943 3907024896  1.8T fd Linux raid autodetect
# fdisk -l /dev/sdf | grep /dev
Disk /dev/sdf: 3.7 TiB, 4000787030016 bytes, 976754646 sectors
/dev/sdf1 256 976754645 976754390  3.7T 8e Linux LVM

If you then put a single LV into the VG which covers the whole VG you
don't benefit much from LVM's functionality, except that you can
easily change allocations later if you decide so.  Re-partitioning is
more complicated.  But even then you have nice and stable device
names.  You could even add or remove drives to the volume group to
extend it, spread logical volumes across the drives and still no LV
name would change.

I like having nice device names like /dev/vg0/root, /dev/vg0/usr,
/dev/vg0/var, /dev/vg0/home, /dev/vg0/swap, /dev/vg0/ for all of
my (currently 4) virtual machines.  And use it a lot, because it so
easy to add/delete/change:

# ls -l /dev/mapper | wc -l
27

For example if I want to test something with btrfs, I can run

lvcreate -n btrfs-test -L 4G vg0

and I have a /dev/vg0/btrfs-test to work with.  No re-partitioning, no
problem with re-reading partition tables which are in use, etc.

urs

Gnus/procmail doesn't read new mails

[Urs Thuermann]

I want to move my mail server and Gnus MUA from a very old machine
(emacs 20.7.1 and gnus 5.8.8) to a Debian bullseye machine with emacs
and Gnus 5.13.

The mail is filtered by procmail into several files in the ~/PROCMAIL
directory.  From there it should be read by Gnus and stored in mail
groups in ~/Mail/.

Filtering to ~/PROCMAIL works as in my old machine:

$ ls -l ~/PROCMAIL
total 12
-rw--- 1 urs urs 709 Nov 10 13:54 bar.gnus
-rw--- 1 urs urs 709 Nov 10 09:13 foo.gnus
-rw--- 1 urs urs 705 Nov 10 09:13 mail.gnus

However, Gnus doesn't read the mails from there.  My .gnus config file
is

$ cat ~/.gnus
(setq gnus-secondary-select-methods '((nnml "")))

(setq mail-sources '((directory :path "~/PROCMAIL/"
:suffix ".gnus")))

When starting Gnus with 'M-x gnus' the .gnus file is read but the
configured mail-sources aren't read.  Running strace on emacs shows
that the PROCMAIL directory isn't touched at all.  However, the (setq
mail-sources...) isn't ignored completely, since mail also is not read
from /var/mail/urs which would be the default.

If I disable procmail and leave incoming mail in /var/mail/urs and if
I remove the (setq mail-sources ...) form .gnus but keep the
gnus-secondary-select-methods, mails are read from /var/mail/urs and
written to ~/Mail/mail/misc as docuemnted in the Gnus manual.

What am I doing wrong that the 'directory' mail source doesn't work?

urs

Re: Gnus/procmail doesn't read new mails

[Urs Thuermann]

Eric S Fraga  writes:

> Just in case, what happens if you expand "~" in the path to PROCMAIL?

That was one of the first things I've tested but that didn't change
anything.

urs

How to check for scheduled shutdown

[Urs Thuermann]

After shutdown -h  I see no way to see this scheduled shutdown.
Before systemd, I could always see the shutdown process with its
arguments using ps(1).

Now, the call to shutdown returns to the shell immediately leaving no
process.  It probably communicates to the init process 1, but, as
usual for systemd, very little or nothing seems to be documented.  Or
at least it's hidden, so that you cannot find it in reasonable time.

I couldn't find any relevant differences with and without a scheduled
shutdown in the output of systemctl status --all and systemctl show --all.

urs

Re: How to check for scheduled shutdown

[Urs Thuermann]

Kamil Joñca  writes:

> kjonca@alfa:~%man shutdown
> SHUTDOWN(8)   
> 
> shutdown  
> 
> SHUTDOWN(8)
> 
> NAME
>shutdown - Halt, power off or reboot the machine
> [...]
> OPTIONS
> [...]
>--show
>Show a pending shutdown action and time if there is any.
> 
> kjonca@alfa:~%sudo shutdown --show
> No scheduled shutdown.
> 
> Am I overlooked something?

On my machine, the man page doesn't mention --show.

$ cat /etc/debian_version 
11.5
$ man shutdown | tail -n1
systemd 247SHUTDOWN(8)
$ /sbin/shutdown --show
/sbin/shutdown: unrecognized option '--show'


urs

Re: pthread_mutex_init man page

[Urs Thuermann]

Riccardo Paolo Bestetti  writes:

> In Debian 9, the
> $ man pthread_mutex_init
> command results in:
> No manual entry for pthread_mutex_init.
> 
> Other pthread man pages (such as pthread_create, pthread_join, ...) are 
> available.
> 
> The "manpages-posix-dev" non-free package is available to install the missing 
> man pages.
> Why are these pages separate from the others, and why are them not free?

Install package glibc-doc.

urs

Re: en_DK.UTF-8 UTF-8

[Urs Thuermann]

info  writes:

> Why can I not fint "en_DK.UTF-8 UTF-8" when I install a normal os from
> you (it is not a server version)??
> 
> I know that I can set it after the installation, but it is not the same!!

You have to generate all locales you want to use.  Run

dpkg-reconfigure locales

and select the required locales.  You can also set the default locale
to en_DK.UTF-8, select the locale at login time or run a single
process in that locale by setting the environment variable LANG, e.g.

LANG=en_DK.UTF-8 ls

You might want to make sure that no other locale environment
variables, i.e. LC_* are set, otherwise they override your LANG
setting.

urs

Re: How to install Debian on a diskless NFS client?

[Urs Thuermann]

Sorry for my late reply.  I have been busy with other things so
continued my diskless install a couple of weeks after your answer.


"[EMAIL PROTECTED]" <[EMAIL PROTECTED]> writes:

> I think what you want to do now is something like this:
> mount -t proc none /tftpboot/client-dir
> chroot /tftpboot/client-dir bin/bash
> 
> Then do basic tasks like;
>  passwd - set a root password (and perhaps adduser a new user)
>  /etc/hostname - set up the hosts name
> 
> [...]

OK, this is the second way to do the install that I have asked in my
first posting.  I did do this now and succeeded.  So thanks for your
answer.

However, I think the installer should be modified to not only support
booting via DHCP/TFTP, but also installing on NFS mounted directories
instead of local disks only.

I see several reason why I think this would be better than doing the
install on the server using debootsrap:
* Installation from the usual install would be easier
* The installation as you suggested requires that the server must also
  be running Linux and be of the same architecture.  Otherwise, you
  couldn't mount /proc and chroot into the directory created by
  debootsrap.

Also, I think the required modifications in the installer would
probably not that large.

urs


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

Xen and PAE

[Urs Thuermann]

I have two separate questions, actually, concerning Xen and PAE:

Today, I installed Xen on a Pentium4 machine running Debian testing.
I did

aptitude install xen-linux-system-2.6.18-4-xen-686

which installed besides others the two debian packages

xen-hypervisor-3.0.3-1-i386-pae and linux-image-2.6.18-4-xen-686

The description of the first of these packages says

This version of the hypervisor is built with PAE enabled, in order
to support systems with more than 4GB of memory.  If you have less
than that you should probably choose the non -pae version.

However, the xen-linux-system-2.6.18-4-xen-686 depends on the -pae
version and the kernel linux-image-2.6.18-4-xen-686 cannot run with
the non -pae xen hypervisor.  An older version of xen-linux-system
seems not be available anymore in Debian testing.

Therefore, I run the -pae version now but I wonder how much
unnecessary overhead and performance impact this might have on my
system with only 512 MB of RAM.  And will there be a version of the
2.6.18-4 kernel for Xen but without -pae?




On another system I had Xen installed for quite a while now.  It had
xen-linux-system-2.6.18-3-xen-686 which depends on the non -pae
hypervisor and kernel found in the packages

xen-hypervisor-3.0.3-1-i386 and xen-linux-system-2.6.18-3-xen-686

Today, I installed additionally xen-linux-system-2.6.18-4-xen-686
because I thought this has a newer kernel and not being aware this has
-pae compiled in.  The installation also called update-grub which then
produced 4 boot entries for Xen/Linux, two of which can't boot
successfully.  The 4 entries are

title   Xen 3.0.3-1-i386-pae / Debian GNU/Linux, kernel 
2.6.18-4-xen-686
...

title   Xen 3.0.3-1-i386 / Debian GNU/Linux, kernel 2.6.18-4-xen-686
...

title   Xen 3.0.3-1-i386-pae / Debian GNU/Linux, kernel 
2.6.18-3-xen-686
...

title   Xen 3.0.3-1-i386 / Debian GNU/Linux, kernel 2.6.18-3-xen-686
...


where the 2nd and the 3rd don't boot because they combine a non-pae
Xen hypervisor with a pae kernel or vice versa.  Even worse, the 2nd
entry was made the default for some reason.

Couldn't grub somehow recognize this problem and leave out those bad
combinations which don't boot?


urs


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

How to install Debian on a diskless NFS client?

[Urs Thuermann]

I want to install Debian testing on a diskless machine (i.e. no
floppy, no CDROM, no hard disk) using DHCP/TFTP and NFS.  Besides
being diskless, the machine is a standard Pentium PC with a NIC which
has an etherboot EPROM to boot via DHCP/TFTP.

As described in the Debian Installation Guide, section 4.6 "Preparing
Files for TFTP Net Booting", I have downloaded and installed
netboot.tar.gz (from etch) on the server machine, configured DHCP and
TFTP (and GRUB, since instead of PXE, I boot nbgrub via TFTP, then
load the debian kernel and initd.gz from GRUB).

I can boot the debian kernel, initrd and debian installer on the
diskless client, choose language, country, keyboard layout, timezone,
configure the clock and network, configure the debian mirror to use.
But then the installer wants to detect and partition disk drives to do
the installation on.  It doesn't allow to select an NFS-mounted
directory to install.  When skipping the partitioning the installer
can't proceed, since it has no disk to use as a target.

I then tried to switch to a shell and to manually do a

   mkdir /target; mount server:/tftpboot/client-dir /target

since I know /target is used in other installations.  However, the
mount command in the initrd's busybox seems to be unable to do NFS
mounts.  This is where I am stuck.  How can I do the installation to
the server disk when the client isn't able to mount it?



A second way would be not to run the installation on the client, but
prepare the directory to be exported to the client directly on the
server.  I think this is what debootsrap is for, but I haven't found
any good documentation how to use it.  The Debian Installation Guide
doesn't describe it.  According to the man page I have run

   ARCH=i386 debootstrap etch /tftpboot/client-dir

which populates the target directory with a minimal system but don't
know how to proceed from here.  I could probably tweak the result by
some editing in /etc, installing/building a kernel and initrd, but I
hope there is an easier way.


Can anyone give me a hint on this?

urs


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

debian testing netinstall via pppoe?

[Urs Thuermann]

Is it possible to use the etch netinstall CD to install a debian
testing system via pppoe?  I have tried but didn't succeed.  The
installer only lets me choose between the two ethernet network cards
and then asks me for IP configuration for the selected NIC.  But I'd
need to configure ppp parameters, i.e. username and password for my
provider.

urs


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

Re: debian testing netinstall via pppoe?

[Urs Thuermann]

Dave Patterson <[EMAIL PROTECTED]> writes:

> Using the netinstall ISO (not the businesscard), do a normal installation,
> leaving the network unconfigured.  This will be a base install only, but
> the package pppoeconf is on that cd.  Once you have the base install in,
> simply apt-get install pppoeconf, and use that to get your connection
> going.  From there, edit your /etc/apt/sources.list, and install whatever
> else you need from there.

OK, thanks.  I have now already installed using the binary-1 CD.  I
have installed quite a couple of Debian systems using netinstall
(indeed all installations I have done were netinstall installations),
but it was always in a working network environment.  This time I
wanted to replace my home brewn completely self compiled
router/mail/news/dhcp/dns server with pppoe to the provider with a
Debian installation.  I still have some problems with the SCSI and
LVM2 setup.  See other posting.

urs


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

Problems booting from SCSI disk (with LVM2)

[Urs Thuermann]

I have installed Debian testing using the binary-1 CD.  After fiddling
with the grub install (see below [*]) the kernel-2.6.15-1-i686-smp
boots and starts from the initramfs.  The initramfs initializes among
other things the SCSI subsystem (AIC7899), device mapper and LVM2, and
my ethernet card (two port adaptec, starfire driver).  But mounting
the real root device fails, because of the order of initializations,
AFAICS.

I see kernel messages for the initialization of the SCSI host
controller, but before this completes, the kernel prints messages for
the initialization of the NIC, then I see a message that mounting
/dev/vg0/root failed, then the messages for the initialization of the
SCSI host controller complete followed ty the kernel message for
recognized partitions on /dev/sda.

It seems that the modprobe for the SCSI host controller starts
correctly before LVM2 activation and mounting the root fs, but the
modprobe seems to return before the SCSI host controller
initialization is complete, so the LVM2 and mount start too early.  I
haven't see this behavior with Linux-2.4 and modutils.  Has this
changed with Linux-2.6 and module-init-tools?  How can I change this?



[*] I had to change the devices.map file and then do a
chroot /target grub-install "(hd0)" on the shell in console 2 to
fix this.
Selecting "Install GRUB boot loader" from the installer always
overwrote my changed devices.map file.  There should be an easier
way to do this.


urs


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

Re: Uniq is not unique ?

[Urs Thuermann]

"Chris Henry" <[EMAIL PROTECTED]> writes:

> Uniq only filters consecutive repeated lines, e.g.
> 
> A
> A
> B
> A
> 
> will become
> 
> A
> B
> A
> 
> If you need it to filter such that only 1 unique line remains, you
> will need to sort first then pipe to uniq (not a good solution for
> really large files).

I sometimes need to filter repeated lines that are not consecutive,
and I use the following simple perl script for this purpose.  Runs
reasonable fast even for large (couple of tens of MB) files:

#!/usr/bin/perl

while (<>) {
if (!$h{$_}) {
$h{$_} = 1;
print;
}
}

HTH,
urs


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

debootstrap and GRUB bootloader

[Urs Thuermann]

I want to build a Debian system using debootstrap and install that on
a Compact Flash card for an x86 embedded system.

I have created and mounted a fresh file system on a logical volume
which I install the new Debian system into using debootstrap.  That
works fine and I can chroot into the new system and install some more
package using aptitude.  This is done by a simple shell script:

#!/bin/sh
mkfs  /dev/vg0/cu
mount /dev/vg0/cu cu
debootstrap lenny cu
cp config/etc/hostname cu/etc
cp config/etc/fstabcu/etc
...
chroot cu <

aptitude update and pdiff files

[Urs Thuermann]

For a couple of month I have seen an annoying change in behavior of
aptitude update and apt-get update.  Formerly, a couple of package
files where downloaded which took a few seconds.  Now I normally see
that aptitude update downloads hundreds of files named like 

Get:282 2006-09-11-1318.15.pdiff [13.3kB]
Get:283 2006-09-11-1318.15.pdiff [13.3kB]
Get:284 2006-09-11-1318.15.pdiff [34.3kB]
Get:285 2006-09-11-1318.15.pdiff [34.3kB]
Get:286 2006-09-11-1318.15.pdiff [34.3kB]

and this takes quite long.  From the name I guess these are only
differences two package files so that less data have to be
transferred.  However, the overhead of nearly 300 TCP connections
causes this to take *much* longer on some systems than simply
downloading the whole package file.

urs


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

Re: aptitude update and pdiff files

[Urs Thuermann]

Jochen Schulz <[EMAIL PROTECTED]> writes:

> Apt is configurable in this respect, it just isn't documented yet (see
> Bug #376158):
> 
> apt-get update -o Acquire::Pdiffs=false
> 
> or, to make this setting permanent, put the following in
> /etc/apt/apt.conf:
> 
> Acquire::PDiffs "false";

Thanks, works much better now.  On machines running 24/7 I will put
the aptitude update in a cron job, on some other the above option will
be more appropriate.

urs


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

what happened to xon?

[Urs Thuermann]

I run Debian testing and have upgraded many packages recently.  After
the upgrade I noticed that the xon(1) utility has gone.  It has been
part of the xutils package before.  That package also contained a
number of other tools which seem to have been moved to other packages
named x11-*, but no package contains xon anymore.

Is this a bug or intended?

urs


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

Debian boxes not replying to multicast and broadcast pings

[Urs Thuermann]

I have currently 4 Linux machines in my network, one embedded system
running Linux-2.4.20, a system build from scractch over the last 13
years running 2.4.34.2 and two Debian boxes, both running Debian's
Linux kernel 2.6.18-4-686.

When I ping to the broacast address or the allhosts multicast address
I get replies from the embedded and the self-built sysem, but not from
the two Debian boxes:

janus:urs$ ping -w2 10.0.0.255
PING 10.0.0.255 (10.0.0.255): 56 octets data
64 octets from 10.0.0.1: icmp_seq=0 ttl=64 time=0.2 ms
64 octets from 10.0.0.31: icmp_seq=0 ttl=64 time=0.7 ms (DUP!)
64 octets from 10.0.0.1: icmp_seq=1 ttl=64 time=0.1 ms
64 octets from 10.0.0.31: icmp_seq=1 ttl=64 time=0.6 ms (DUP!)

--- 10.0.0.255 ping statistics ---
2 packets transmitted, 2 packets received, +2 duplicates, 0% packet loss
round-trip min/avg/max = 0.1/0.4/0.7 ms
janus:urs$ ping -w2 224.0.0.1 
PING 224.0.0.1 (224.0.0.1): 56 octets data
64 octets from 10.0.0.1: icmp_seq=0 ttl=64 time=0.3 ms
64 octets from 10.0.0.31: icmp_seq=0 ttl=64 time=0.9 ms (DUP!)
64 octets from 10.0.0.1: icmp_seq=1 ttl=64 time=0.1 ms
64 octets from 10.0.0.31: icmp_seq=1 ttl=64 time=0.6 ms (DUP!)

--- 224.0.0.1 ping statistics ---
2 packets transmitted, 2 packets received, +2 duplicates, 0% packet loss
round-trip min/avg/max = 0.1/0.4/0.9 ms

The two Debian machines do reply to unicast pings, however:

janus:urs$ ping -c1 blue
PING blue.isnogud.escape.de (10.0.0.2): 56 octets data
64 octets from 10.0.0.2: icmp_seq=0 ttl=64 time=1.6 ms

--- blue.isnogud.escape.de ping statistics ---
1 packets transmitted, 1 packets received, 0% packet loss
round-trip min/avg/max = 1.6/1.6/1.6 ms
janus:urs$ ping -c1 mop 
PING mop.isnogud.escape.de (10.0.0.65): 56 octets data
64 octets from 10.0.0.65: icmp_seq=0 ttl=64 time=0.3 ms

--- mop.isnogud.escape.de ping statistics ---
1 packets transmitted, 1 packets received, 0% packet loss
round-trip min/avg/max = 0.3/0.3/0.3 ms

Also, I have written two small test programs, udp-ping and udp-echo,
which send short packets two a specified UDP port and echo all packets
to a certain UDP port, respectively.  If I send UDP packets to the
multicast address, udp-echo will answer, also on the Debian machines.
This shows that the Debian machines are able to receive multicast.

So what is the reason, that the 2.6.18-4-686 kernel does not reply to
ICMP echo request to broadcast and multicast address?  I've looked
through /proc/net and /proc/sys/net but haven't found anything which
looked relevant.

urs


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

Re: Debian boxes not replying to multicast and broadcast pings

[Urs Thuermann]

Jeff D <[EMAIL PROTECTED]> writes:

> if you want it to reply to broadcast pings:
> echo 0 > /proc/sys/net/ipv4/icmp_echo_ignore_broadcasts

Ah, thanks.  This not only causes replies to broadcast pings to be
sent but also multicast to 224.0.0.1 (which is allhosts multicast and
essentially the same as broadcast).

> but, its turned off by default for security reasons, to avoid smurf
> attacks, which will try to flood a host with icmp ping requests.

When I looked yesterday, I've seen that file and I seem to remember it
already contained 0.  Today it was 1, so probably my memory on this is
wrong.

Also, I may have not looked that much at icmp_echo_ignore_broadcasts,
since I was actually more after multicast.  The ping to broadcast was
just a additional test.


urs


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

Which software to copy and edit video from MiniDV camcorder

[Urs Thuermann]

I want to copy video recordings from my MiniDV camcorder to a Linux
machine, edit it, convert it to MPEG2, and burn it to DVD.  What
software is available in Debian for these tasks.  I have read
something about kino, installed it, but it seems to be quite
instable.

After starting and clicking on "Capture", kino crashes in >90% of all
cases with the following message (I have also included the last lines
of kino output on its terminal):

>>> Using iec61883 capture
>>> iec61883Reader::StartThread on port 0
>> Trying XVideo at 720x576
>>> XvQueryAdaptors count: 1
>>> Xv: ATI Radeon Video Overlay: ports 65 - 65
>>> formats supported: 8
>>> 0x41424752 (RGBA) packed
>>> 0x0 () packed
>>> 0x54424752 (RGBT) packed
>>> 0x32424752 (RGB2) packed
>>> 0x32595559 (YUY2) packed
>>> 0x59565955 (UYVY) packed
>>> 0x32315659 (YV12) planar
>>> 0x30323449 (I420) planar
>>> 0: XV_IMAGE, 2048x2048 rate = 1/1
The program 'kino' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadAccess (attempt to access private resource denied)'.
  (Details: serial 6520 error_code 10 request_code 143 minor_code 1)
  (Note to programmers: normally, X errors are reported asynchronously;
   that is, you will receive the error a while after causing it.
   To debug your program, run it with the --sync command line
   option to change this behavior. You can then get a meaningful
   backtrace from your debugger if you break on the gdk_x_error() function.)

A few times I was able to get past this point and control the camera,
i.e. I could click "play" and "stop" and the camera reacted
accordingly, I was also able to capture some scenes.  However, rewind
and fast forward did not work.  Also, in most cases kino will stop at
some point with a message, similar to the above.

Have other users similar experiences or are there other packages which
I should look at?

urs


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

Re: Which software to copy and edit video from MiniDV camcorder

[Urs Thuermann]

Ron Johnson <[EMAIL PROTECTED]> writes:

> You don't tell us which branch (Sarge, Etch, testing, Sid) and
> version of kino that you are running.

Yes, sorry.  I use kino 0.92-3 on Debian testing.  I also forgot to
mention another detail, which might be the cause for my problems[1]:
The kino and Xserver run on different machines  The Xserver is also
Debian testing with current xorg.

In the meantime I also found out, that I can prevent kino from running
into this error, if I change "Display" in the Preferences from XVideo
to GDK.

> Also, there is a kino web site and forums that should be of more
> assistance to you.

Thanks for that hint.


urs

[1] It's a pity that there are so many programmers coming to linux
with no unix background, who reinvent tools exisiting in unix for
decades, who don't use existing libs, and who have a single user,
personal computer in mind when designing and coding their software
instead of a heterogenous, networked, multi-user environment,
where people run processes distributed over machines and files
systems, running remote shells and X clients.


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

Re: Which software to copy and edit video from MiniDV camcorder

[Urs Thuermann]

Thomas Jollans <[EMAIL PROTECTED]> writes:

> copy: dvgrab
> edit: cinelerra
> ->DVD: mencoder, transcode, dvdauthor. Gentoo forums have a nice howto.

OK, thanks.  I'll look at these.

urs


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

tcpdump doesn't show VLAN IDs

[Urs Thuermann]

tcpdump ion current Debian testing does not show the VLAN ID in 802.1q
tagged Ethernet frames.

I have observed this using two machines:
(A) Linux-2.4.34.4, almost everything compiled on my own from scratch
tcpdump-3.9.5 and libpcap 0.9.5
(B) Debian testing, up to date, kernel 2.6.18-4-686, tcpdump 3.9.5 and
libpcap 0.9.5.

I have configured VLAN 100 on both machines using vconfig add eth0 100, 
have set addresses 172.16.6.1/24 and 172.16.6.2/24 resp. to eth0.100
interfaces, and have set the interfaces up.  The VLAN works.

But if I run tcpdump on eth0 on the Debian machine, it doesn't show me
the VLAN ID.

On (A) I do

   ping -c1 172.16.6.2

and I run tcpdump on both machines on interface eth0:

host-A # tcpdump -ne -i eth0 -xx not port 22
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes
13:55:48.128953 00:00:d1:9d:7b:a8 > 00:90:27:8f:dc:65, ethertype 802.1Q 
(0x8100), length 102: vlan 100, p 0, ethertype IPv4, 172.16.6.1 > 172.16.6.2: 
ICMP echo request, id 44817, seq 0, length 64
0x:  0090 278f dc65  d19d 7ba8 8100 0064
0x0010:  0800 4500 0054  4000 4001 d685 ac10
0x0020:  0601 ac10 0602 0800 dd63 af11  444f
0x0030:  6546 d5f1 0100 0809 0a0b 0c0d 0e0f 1011
0x0040:  1213 1415 1617 1819 1a1b 1c1d 1e1f 2021
0x0050:  2223 2425 2627 2829 2a2b 2c2d 2e2f 3031
13:55:48.129669 00:90:27:8f:dc:65 > 00:00:d1:9d:7b:a8, ethertype 802.1Q 
(0x8100), length 102: vlan 100, p 0, ethertype IPv4, 172.16.6.2 > 172.16.6.1: 
ICMP echo reply, id 44817, seq 0, length 64
0x:   d19d 7ba8 0090 278f dc65 8100 0064
0x0010:  0800 4500 0054 d656  4001 402f ac10
0x0020:  0602 ac10 0601  e563 af11  444f
0x0030:  6546 d5f1 0100 0809 0a0b 0c0d 0e0f 1011
0x0040:  1213 1415 1617 1819 1a1b 1c1d 1e1f 2021
0x0050:  2223 2425 2627 2829 2a2b 2c2d 2e2f 3031

host-B (Debian) # tcpdump -ne -xx -i eth0 not port 22
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes
13:55:48.127450 00:00:d1:9d:7b:a8 > 00:90:27:8f:dc:65, ethertype 802.1Q 
(0x8100), length 102: ethertype IPv4, 172.16.6.1 > 172.16.6.2: ICMP echo 
request, id 44817, seq 0, length 64
0x:  0090 278f dc65  d19d 7ba8 8100 0064
   ^
0x0010:  0800 4500 0054  4000 4001 d685 ac10
0x0020:  0601 ac10 0602 0800 dd63 af11  444f
0x0030:  6546 d5f1 0100 0809 0a0b 0c0d 0e0f 1011
0x0040:  1213 1415 1617 1819 1a1b 1c1d 1e1f 2021
0x0050:  2223 2425 2627 2829 2a2b 2c2d 2e2f 3031
13:55:48.127546 00:90:27:8f:dc:65 > 00:00:d1:9d:7b:a8, ethertype 802.1Q 
(0x8100), length 102: ethertype IPv4, 172.16.6.2 > 172.16.6.1: ICMP echo reply, 
id 44817, seq 0, length 64
0x:   d19d 7ba8 0090 278f dc65 8100 0064
   ^
0x0010:  0800 4500 0054 d656  4001 402f ac10
0x0020:  0602 ac10 0601  e563 af11  444f
0x0030:  6546 d5f1 0100 0809 0a0b 0c0d 0e0f 1011
0x0040:  1213 1415 1617 1819 1a1b 1c1d 1e1f 2021
0x0050:  2223 2425 2627 2829 2a2b 2c2d 2e2f 3031

As you can see, the self-compiled tcpdump shows the VLAN tag, i.e.
"vlan 100, p 0" while the Debian version does not, although it sees
the complete ethernet frame header including the VLAN tag (marked
with ^ in the hexdump) and it shows that it is a VLAN tagged
frame.

Has Debian patched the tcpdump src or is this a bug?

urs


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

package management begins to annoy me

[Urs Thuermann]

I use Debian testing on 2 desktop machines and a notebook, the oldest
of them is 4-5 years old.  While in the begining I found apt-get and
dpkg quite usable (but didn't like dselect), now aptitude tends more
and more to annoy me, for several reasons.  Maybe, and I hope so, this
is only because i don't know apt-get and aptitude well enough.

1. aptitude has the nice feature of marking packages that are install
   automatically, qhich I always missed in apt-get.  But every once in
   a while I check the installed package with

aptitude search . | grep ^i

   (BTW, is there a simpler way to do this?) and I quite often see
   installed packages, which are not marked 'A' for automatically
   installed but which I definitely know I have never installed
   manually.  Unfortunately, I currently don't have examples at hand.

2. Every 1-2 months or so I do a

aptitude update && aptitude -R safe-upgrade

   but more and more often I see aptitude wanting to bloat my
   installation of currently roughly 2 GB by another 400 or 500 MB by
   installing hundreds of new packages.  My suspect is that
   increasingly many packages have broken dependencies and want to
   pull in quite a lot of other packages which they really shouldn't
   depend on.

   However, I don't know for sure concrete examples for this, except
   that 2+ years ago, I wanted to upgrade the already installed CD
   ripper grip by running

aptitude install grip

   and this insisted on installing almost the complete cups system.  I
   found this completely broken since I don't had a printer at all and
   you can use grip quite well without printing.

3. On a Debian testing system at work, where I haven't upgraded for
   maybe 3-4 months I ran

aptitude -R safe-upgrade

   which caused aptitude to run for an hour generating thousands of
   messages about resolving open/closed/defered dependency conflicts
   and then giving up.  I was only able to upgrade package for package
   explicitly for a couple dozen packages, then safe-upgrade worked
   again.

4. On my notebook I have today safe-upgraded with -R (which caused an
   increase of 140 MB to beof the installed size) and now the system
   seems to be quite up-to-date.  Nevertheless, I now have problems
   installing new packages involving perl (like I have had several
   times before).  aptitude wants to upgrade perl, perl-base, and
   perl-modules, then it detects some unmet dependencies

The following packages have unmet dependencies:
  libglib-perl: Depends: perlapi-5.8.8 which is a virtual package.
  libcompress-zlib-perl: Depends: perlapi-5.8.8 which is a virtual 
package.
  libperl5.8: Depends: perl-base (= 5.8.8-12) but 5.10.0-19 is to be 
installed.
  ...

   which it wants to resolve by removing

Remove the following packages:
abiword-gnome
cogito
git-core
gnome
gnome-office
libcompress-zlib-perl
libdigest-sha1-perl
libft-perl
libperl5.8

   which I don't want to accept but all other following suggestions
   aren't better.  In the past, upgrading perl has also caused svk to
   be removed, and I wasn't able to reinstall svk although I really
   missed it.

5. Doing an

aptitude full-upgrade

   seems to solve these problems with perl in 4. and upgrades perl to
   the current version but it also wants to install 327 new packages
   increasing disk usage by 675 MB:

76 packages upgraded, 327 newly installed, 21 to remove and 1 not 
upgraded.
Need to get 432MB of archives. After unpacking 675MB will be used.

   which find quite expensive since I only want to upgrade, not
   install a whole lot of unneeded packages.  With apt-get it looks
   the same.

So, am I doing something completely wrong here?


urs


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

wpa_supplicant and dhclient

[Urs Thuermann]

I want to connect an embeded mobile WLAN station to a WLAN network
with WPA-PSK and DHCP.  In /etc/network/interfaces I have

auto ath0
iface ath0 inet dhcp
wpa-driver wext
wpa-conf /etc/wpa_supplicant/ww.conf

My problem is when I boot while the WLAN is not available.  Then ifup
automatically starts wpa_supplicant and dhclient to set up the
interface.  But dhclient will not be able to get a configuration by
DHCP and waits for a number of intervalls which causes quite a long
delay in boot time.  Then ifup returns and the boot process continues,
while wpa_supplicant and dhclient continue to work in the background.

The second problem is this: When at some time the WLAN becomes
available, wpa_supplicant will establish the connection to the access
point but it doesn't inform dhclient to send a new DHCP request
immediately.  Instead, dhclient sends it DHCP request in very long,
randomized intervalls, so it can take very long (like 20 or 30
minutes) until the interface is configured from the time the WLAN
becomes available.

Therefore, my questions are:

1. Is it possible to configure ifup so that it goes to backgound
   immediately so that the boot process isn't delayed?

2. Can wpa_supplicant and dhclient talk to each other so that dhclient
   doesn't start to send DHCP requests before the WPA connection is
   established but then after connection sends its requests in short
   intervalls?

urs


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

Re: wpa_supplicant and dhclient

[Urs Thuermann]

Andrei Popescu <[EMAIL PROTECTED]> writes:

> Try setting 'CONCURRENCY=startpar' in /etc/init.d/rc, but do read the 
> notes before.

Interesting to know, but I didn't need this, since your other
suggestion below also solves the boot delay.  Using wpa-roam, ifup
returns quickly and the boot process continues.

> I think the roaming mode might be better suited for your needs. See 
> README.Debian for more info.

Thanx for your answer.  I wasn't aware of wpa-roam and wpa_action at
all.  Now I have tried it and it does exactly what I was looking for.
Now, I need some time to read all the shell scripts, wpa_cli scripts,
etc. to understand exactly what's going on.


urs


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

Re: wpa_supplicant and dhclient

[Urs Thuermann]

Dominique Dumont <[EMAIL PROTECTED]> writes:

> >> No. DHCP is not accepted in roaming mode.
> >
> > Where did you get that from
> 
> I tried and I got an explicit error message ;-)
> 
> Only manual mode is accepted.

As Andrei wrote, read
/usr/share/doc/wpa_supplicant/README.Debian.  There you'll find
everything explained how to use roaming with DHCP or static addresses
and more.

urs


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

Prevent aptitude from starting services

[Urs Thuermann]

Is it possible to prevent aptitude from automatically running services
after install?  That is, when I install some daemon or some other
package with a start script in /etc/init.d, e.g. bind or ferm, I don't
want that init script to be executed.  This is annoying when services
are started before I have configured them, and also if I build a
system with debootstrap, then chroot into the new system to install
some further packages and then services are run or netfilter rules are
changed on the build system.

Looking through the man page of aptitude I haven't found an option to
do this.


urs


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

Re: cp, but preserve the dest attribute

[Urs Thuermann]

T o n g  writes:

> cp -a, or rsync -a, or cpio normally preserve the source attribute and 
> set the destination attribute accordingly. Now my question is, is there 
> any way for me to preserve the destination attribute and disregard the 
> source attribute. Any way to make it possible, cp/rsync/cpio or something 
> else?

touch -r dst /tmp/ref && cp src dst && touch -r /tmp/ref dst

urs


-- 
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/ygftxsk38a0@janus.isnogud.escape.de

No static libmpg123.a

[Urs Thuermann]

The changelog for libmpg123-dev says:

mpg123 (1.13.7-3) experimental; urgency=low

  * Enable LFS aliases, attempt to fix FTBFS on kfreebsd-i386.
  * Drop static library.
  * Improve ARM CPUs optimizations.

 -- Alessio Treglia   Thu, 29 Mar 2012 14:13:02
+0200

Why has the static library been dropped?

urs


-- 
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/ygfobis37tj@janus.isnogud.escape.de

Re: time zone and UTC issue

[Urs Thuermann]

"J. B"  writes:

> My box is configured to the local time zone from beginning, both
> hwclock and system time.  But linux always favor hwclock to
> UTC. What is the advantage of doing that ?

Although time, timezones and clock setting are quite a simple topic it
seems to be major source of confusion for many people and many systems
are mis-configured, as can seen in this thread and as can be often
seen in E-mail time stamps.

I jumped into this thread somewhat late and wanted to write a couple
of sentences, but now this has become a little bit longer than
intended.

First, to reduce confusion, it is helpful to use good terminology and
to know what the meaning of used terms is.

1. Local time and UTC.

   I strongly dislike the term "local time".  There is no such thing
   as a local time or different local times in different locations (on
   earth).  There is only one global time.  What differs locally is
   only the /represantation/ of the time.  Nevertheless, I also
   sometimes use "local time" to mean "local time representation".
   Also, UTC is only /one/ possible representation for times.  It is
   special in that is "universal", i.e. it is used independent of
   location and doesn't have weird things like daylight saving times.
   Both, UTC and local time representation, express time as year,
   month, day, hour, minute, second, and (hopefully) timezone
   indicator.

2. System clock and system time.

   The system clock is a software clock driven by the Linux kernel and
   it defines the system time.  The system clock uses yet another time
   representation.  Although seemingly everybody states that Linux and
   Unix systems run their system clock in UTC, this is IMO not quite
   correct.  The system time, time stamps in Unix file systems, and
   time stamps exchanged between the Linux kernel and user space
   programs are representated as a POSIX time_t (or struct timeval or
   struct timespec).  This representation is only an integer number
   counting the seconds since the POSIX epoch (ignoring leap seconds).
   The timeval and timespec representation additionally give the
   microseconds and nanoseconds, respectively.  The POSIX epoch is a
   fixed point in time, usually given in UTC time representation and
   it is January 1, 1970 0:00:00 UTC.  The representation as a simple
   integer has several advantages:

   * It's very simple to advance the clock by one second.  No need to
 carry to the next unit after 60 seconds, 60 minutes, 24 hours,
 and so on.
   
   * It's very simple to calculate the difference between two times.
   
   * It's independent of local time representation and daylight saving
 rules.  It doesn't jump forth and back.
   
   When people say "The Linux kernel system time is always in UTC",
   what the really mean is that the kernel keeps a time representation
   that doesn't depend on any local timezone definitions, doesn't jump
   for daylight saving times and is based on a certain point in time
   (the epoch) which is usually expressed in UTC (although you could
   equally well define the epoch as "December 31, 1969 19:00:00
   Eastern Standard Time" because that is the /same/ time as
   1970-01-01 0:00:00 UTC).
   
   You can get the current kernel system time with the gettimeofday(2)
   or clock_gettime(2) system calls, or print it like this
   
   $ date +%s
   1354558752
   $ perl -e 'print time,"\n"'
   1354558754

   I think it's important to understand that you don't set the system
   clock to one time or another, you don't set it to "local time" nor
   to UTC.  You set it to "the current time" (or have it synchronized
   to "the current time" using NTP), since there is only one global
   time.  The representation used by the system clock is POSIX struct
   timespec and nothing else.

3. Hardware clock (aka. CMOS clock, BIOS clock or real time clock (RTC)).

   This is a clock that ticks independently from any operating system
   and even keeps running when the computer is turned off.  This clock
   represents its time as year-month-day-weekday-hour-minute-second so
   it can be set to any local time representation or to UTC.
   Unfortunately, it doesn't have information which representation it
   is set to nor if it currently is set to daylight saving time or
   not.  Without this information, you have to guess the meaning of
   the time you read from the hardware clock.  Therefore, it's best
   best to use UTC for this clock since that never has be adjusted for
   daylight saving time.  Otherwise, in multi-boot system each OS may
   adjust thue RTC by one hour resulting in wrong RTC time.

   This clock is normally not used in your Linux system, except when
   booting the system.  Early in the boot process the time is read
   from the hardware clock, its time representation is
   interpreted/guess (hopefully correct), the POSIX time_t calculated
   and the system time set from this.  After that, you can read and
   set the hardware clo

Re: time zone and UTC issue [rant]

[Urs Thuermann]

Ralf Mardorf  writes:

> If I save BIOS settings as a file and the hwclock is set to UTC, the
> files don't get the German time. The BIOS is the BIOS, it's neither
> Windows, I don't use Windows, but nor the BIOS is Linux, so Linux can't
> "translate" UTC to local time, when I save BIOS settings.

The problem is that you store to a file system that is broken the same
way as the hardware clock, i.e. the MSDOS or VFAT file system that has
no way to indicate what timezone the file time stamps should be
interpreted in.

If you access the vfat file system on your floppy/USB stick using
mtools, you can run the mtools commands in the appropriate TZ
environment.  mcopy -m from vfat to linux file system
(ext2/3/4,xfs,btrfs,...) will create the destination file with the
correct POSIX time stamp.


urs


-- 
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/ygf1uf7klzq@janus.isnogud.escape.de

Re: time zone and UTC issue [rant]

[Urs Thuermann]

Ralf Mardorf  writes:

> If the clock does use local time, then the time for all BIOS and all
> Linux files are ok.

This is not completely true.  If there is change from/to daylight
saving time to/from standard time between saving the files using the
BIOS and booting your Linux system the kernel will see wrong time
stamp on VFAT file systems.

When mounting a vfat file system the kernel will interpret the time
stamps in the file system according to the timezone set using
settimeofday(2).  This is usually set only once when the system boots
and it is set to current time offset in minutes compared to UTC.  If
this doesn't match the offset when the file were originally created on
the vfat file system, time stamps will be interpreted wrong.  I see
this often when I copy pictures from my digital camera's SD card.

IMO this is a flaw in the kernel.  The kernel should never use the
timezone set with settimeofday().  If should either have a more
sophisticated timezone handling or not use timezones at all.

A couple of years ago I have begun working on a kernel implementation
for full timezone handling and it worked.  But it was for linux 2.4
and the code was quite a quick hack.


urs


-- 
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/ygfwqwzj6oh@janus.isnogud.escape.de

Re: time zone and UTC issue [rant]

[Urs Thuermann]

Ralf Mardorf  writes:

> spinymouse@q:~$ ls -l /media/spinymouse/INTENSO/
> total 32
> -rw-r--r-- 1 spinymouse spinymouse   304 Oct 22  2011 B22OCT11.CMO
> -rw-r--r-- 1 spinymouse spinymouse   304 Sep 30  2011 B30SEP11.CMO
> -rw-r--r-- 1 spinymouse spinymouse 15644 Nov 10  2011 Hakle-Geld-zurück.odt
> -rw-r--r-- 1 spinymouse spinymouse33 Oct 13 23:11 
> Ubuntu_Studio_USB_stick_test.txt
> -rw-r--r-- 1 spinymouse spinymouse30 Oct 13 23:11 
> Ubuntu_Studio_USB_stick_test.txt~
> 
> Nautilus does show time and date for all files of the FAT formatted USB
> stick. I'm not sure if time and date will differ, if I switch local and
> I won't test it now. Only the first two files are from the BIOS, the
> other 3 files are from Linux installs.

There seems to be another wrong assumption here.  Both, ls and
nautilus can show the time (not only date) of all files.  By default,
ls -l shows the date including the year if the file is older than 6
months, and the date without the year but with time, otherwise.

With GNU ls, you can use --time-style=long-iso to get minutes and
--time-style=full-iso to get the full time stamp including even
seconds and nano seconds for each file.

This has nothing to with the type of file system the file resides on.
The kernel hides the details of file systems, e.g. the time stamp
representation.  If offers the same interface for all file systems,
for ext2/3/4, vfat, iso9660, and all other file system types.
In this case, it is stat(2) defined by POSIX, which gives the file
time stamp as a POSIX time_t, i.e. in seconds.  Linux and GNU ls use
an extension that gives the time as POSIX struct timespec, i.e. in
seconds and nano seconds.

urs


-- 
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/ygfsj7mkj4c@janus.isnogud.escape.de

Re: [OT] End of the world

[Urs Thuermann]

darkestkhan  writes:

> Isn't this a bit too fast ? Wasn't End of the World scheduled in
> crontab for 21.12.2012 ? ;)

You cannot specify a year using crontab(1) or in /etc/crontab.  You
could, however, specify that the end of the world will happen on each
December 21, in any year :-)

urs


-- 
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/ygfobiakge3@janus.isnogud.escape.de

Re: Why the 64 bit ISA is better

[Urs Thuermann]

Stan Hoeppner  writes:

> Given the recent threads regarding 32 vs 64 bit I thought I'd take a
> moment to present information often omitted in responses to these posts.
> 
> [ some good explanations ]

Another often underestimated bit is the new addressing mode relative
to the instruction pointer %rip.  This greatly reduces the performance
overhead of position independent code needed for shared libs.

On the other hand, the penalties that come with 64 bit are also often
underestimated: Pointers are 64 bit wide and this can increase memory
foot print considerably.  Argueing with decreasing costs for main
memory neglects the fact that even this larger memory must be
inited/loaded from disk, transferred to/from CPU thru the memory bus
and pointers also require more space in CPU cache memory which cannot
easily be enlarged.

Nevertheless, I always prefer the amd64 installation instead of i386
when installing on new machines.

urs


-- 
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/ygfmwu26fmd@janus.isnogud.escape.de

Decoding QR

[Urs Thuermann]

I have searched the Debian repository for packages to decode QR codes
but haven't found anything suitable.

I often take photos of QR codes e.g. with my SLR camera, saving them
as JPG, that I want to decode later on my Debian machine.

I'd like to have a simple command line tool that I can call with a JPG
image file as argument and that searches the image for a QR code,
decodes it and prints the decoded text to stdout.

I have tried with libdecodeqr-examples and libdecodeqr0 but that seems
to be only able to decode the example JPG file from that package.  All
of my images from the SLR were not recognized.

I haven't installed libdecodeqr-dev since that seems to have
completely ill dependencies.  It wants to install 50 new libs of 64 MB
size, ranging from font libs, to firewire communication, computer
vision (openCV), lots of X11 libs, XML, regexps, and lots of other
stuff, that haven't much to do with QR codes.  I don't want to capture
from image devices, I don't want to display QR codes, I don't want any
GUI, X11.  Just decode QR codes in JPG files.

Isn't there anything in Debian for that purpose?

urs


-- 
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/ygftxo73g09@janus.isnogud.escape.de

Re: Decoding QR

[Urs Thuermann]

Claudius Hubig  writes:

> An example image would be helpful.

I have put some to my home page, see below.

> > I haven't installed libdecodeqr-dev since that seems to have
> > completely ill dependencies.
> 
> It’s a development library (-dev) intended for use if you want to,
> well, develop your own programmes.

Yes, I know and I wouldn't mind doing so.  But I didn't like the
dependencies on capturing, font libs, displaying on X11 etc.  IMO a
library for decoding should do exactly and only that.  It shouldn't
mess with where I get the images from or whether I want to display
them.

> zbarimg from zbar-tools was able to recognise [0], but not [1].

Ah, thanks for that hint.  I haven't found that tool before.  I tried
zbarimg on my images but (first) without success.  The images were
taken from a laptop screen and have probably to high a resolution,
showing too much detail of the screen's pixel structure instead of
solid white.  After converting to 320 x 200 pixels with ImageMagick
zbarimg found the QR codes in all images immediately.

I have put my example images on

http://thuermann.net/urs/qr 

The images eos-*.jpg on couldn't be decoded by zbarimg, but
small.eos-*.jpg were decoded successfully.

Thanks again.  That brings me close to my goal of automatically
correcting image time stamps, i.e. to compensate for the camera
clock's time drift.


urs


-- 
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/ygfppyv3axx@janus.isnogud.escape.de

Strange load average

[Urs Thuermann]

I run Debian testing/amd64 with current kernel on an Intel Core2Duo:

$ uname -a
Linux bit 3.2.0-4-amd64 #1 SMP Debian 3.2.41-2 x86_64 GNU/Linux

When the system is idle even for long periods, the 5-minute load
average and 15-minute load average don't drop:

$ while sleep 10; do uptime; done
 09:38:54 up 1 day,  2:03,  4 users,  load average: 0.00, 0.01, 0.05
 09:39:04 up 1 day,  2:03,  4 users,  load average: 0.00, 0.01, 0.05
 09:39:14 up 1 day,  2:03,  4 users,  load average: 0.00, 0.01, 0.05
 09:39:24 up 1 day,  2:03,  4 users,  load average: 0.00, 0.01, 0.05
 09:39:34 up 1 day,  2:04,  4 users,  load average: 0.00, 0.01, 0.05
 09:39:44 up 1 day,  2:04,  4 users,  load average: 0.00, 0.01, 0.05
 09:39:54 up 1 day,  2:04,  4 users,  load average: 0.00, 0.01, 0.05
 09:40:04 up 1 day,  2:04,  4 users,  load average: 0.00, 0.01, 0.05
 09:40:14 up 1 day,  2:04,  4 users,  load average: 0.00, 0.01, 0.05
 09:40:24 up 1 day,  2:04,  4 users,  load average: 0.00, 0.01, 0.05
 09:40:34 up 1 day,  2:05,  4 users,  load average: 0.00, 0.01, 0.05
 09:40:44 up 1 day,  2:05,  4 users,  load average: 0.00, 0.01, 0.05
 09:40:54 up 1 day,  2:05,  4 users,  load average: 0.00, 0.01, 0.05
 09:41:04 up 1 day,  2:05,  4 users,  load average: 0.00, 0.01, 0.05
 09:41:14 up 1 day,  2:05,  4 users,  load average: 0.00, 0.01, 0.05
 09:41:24 up 1 day,  2:05,  4 users,  load average: 0.00, 0.01, 0.05
 09:41:34 up 1 day,  2:06,  4 users,  load average: 0.00, 0.01, 0.05
 09:41:44 up 1 day,  2:06,  4 users,  load average: 0.00, 0.01, 0.05
 09:41:54 up 1 day,  2:06,  4 users,  load average: 0.00, 0.01, 0.05
 09:42:04 up 1 day,  2:06,  4 users,  load average: 0.00, 0.01, 0.05
 09:42:14 up 1 day,  2:06,  4 users,  load average: 0.00, 0.01, 0.05
 09:42:24 up 1 day,  2:06,  4 users,  load average: 0.00, 0.01, 0.05
 09:42:34 up 1 day,  2:07,  4 users,  load average: 0.00, 0.01, 0.05
 09:42:44 up 1 day,  2:07,  4 users,  load average: 0.00, 0.01, 0.05
 09:42:54 up 1 day,  2:07,  4 users,  load average: 0.00, 0.01, 0.05
 09:43:04 up 1 day,  2:07,  4 users,  load average: 0.00, 0.01, 0.05
 09:43:14 up 1 day,  2:07,  4 users,  load average: 0.00, 0.01, 0.05
 09:43:24 up 1 day,  2:07,  4 users,  load average: 0.00, 0.01, 0.05
 09:43:34 up 1 day,  2:08,  4 users,  load average: 0.00, 0.01, 0.05
 09:43:44 up 1 day,  2:08,  4 users,  load average: 0.00, 0.01, 0.05
 09:43:54 up 1 day,  2:08,  4 users,  load average: 0.00, 0.01, 0.05
 09:44:04 up 1 day,  2:08,  4 users,  load average: 0.00, 0.01, 0.05
 09:44:14 up 1 day,  2:08,  4 users,  load average: 0.00, 0.01, 0.05
 09:44:24 up 1 day,  2:08,  4 users,  load average: 0.00, 0.01, 0.05
 09:44:34 up 1 day,  2:09,  4 users,  load average: 0.00, 0.01, 0.05
 09:44:44 up 1 day,  2:09,  4 users,  load average: 0.00, 0.01, 0.05
 09:44:54 up 1 day,  2:09,  4 users,  load average: 0.00, 0.01, 0.05
 09:45:04 up 1 day,  2:09,  4 users,  load average: 0.00, 0.01, 0.05
 09:45:14 up 1 day,  2:09,  4 users,  load average: 0.00, 0.01, 0.05
 09:45:24 up 1 day,  2:09,  4 users,  load average: 0.00, 0.01, 0.05
 09:45:34 up 1 day,  2:10,  4 users,  load average: 0.00, 0.01, 0.05
 09:45:44 up 1 day,  2:10,  4 users,  load average: 0.00, 0.01, 0.05
 09:45:54 up 1 day,  2:10,  4 users,  load average: 0.00, 0.01, 0.05
 09:46:04 up 1 day,  2:10,  4 users,  load average: 0.00, 0.01, 0.05
 09:46:14 up 1 day,  2:10,  4 users,  load average: 0.00, 0.01, 0.05
 09:46:24 up 1 day,  2:10,  4 users,  load average: 0.00, 0.01, 0.05
 09:46:34 up 1 day,  2:11,  4 users,  load average: 0.00, 0.01, 0.05
 09:46:44 up 1 day,  2:11,  4 users,  load average: 0.00, 0.01, 0.05
 09:46:54 up 1 day,  2:11,  4 users,  load average: 0.00, 0.01, 0.05
 09:47:04 up 1 day,  2:11,  4 users,  load average: 0.00, 0.01, 0.05
 09:47:14 up 1 day,  2:11,  4 users,  load average: 0.00, 0.01, 0.05
 09:47:24 up 1 day,  2:11,  4 users,  load average: 0.00, 0.01, 0.05
 09:47:34 up 1 day,  2:12,  4 users,  load average: 0.00, 0.01, 0.05
 09:47:44 up 1 day,  2:12,  4 users,  load average: 0.00, 0.01, 0.05
 09:47:54 up 1 day,  2:12,  4 users,  load average: 0.00, 0.01, 0.05
 09:48:04 up 1 day,  2:12,  4 users,  load average: 0.00, 0.01, 0.05
 09:48:14 up 1 day,  2:12,  4 users,  load average: 0.00, 0.01, 0.05
 09:48:24 up 1 day,  2:12,  4 users,  load average: 0.00, 0.01, 0.05
 09:48:34 up 1 day,  2:13,  4 users,  load average: 0.00, 0.01, 0.05
 09:48:44 up 1 day,  2:13,  4 users,  load average: 0.00, 0.01, 0.05
 09:48:54 up 1 day,  2:13,  4 users,  load average: 0.00, 0.01, 0.05
 09:49:04 up 1 day,  2:13,  4 users,  load average: 0.00, 0.01, 0.05
 09:49:14 up 1 day,  2:13,  4 users,  load average: 0.00, 0.01, 0.05
 09:49:24 up 1 day,  2:13,  4 users,  load average: 0.00, 0.01, 0.05
 09:49:34 up 1 day

Re: Strange load average

[Urs Thuermann]

Darac Marjal  writes:

> Probably not. It's probably just a rounding/display issue.

Rounding issues cannot explain the numbers.  When the load average for
the last minute, sampled every 10s, is 0.00 a each sample, it cannot
be higher when averaged over a longer time.

> Remember that you're asking the system do do some work every ten
> seconds so, when it reads "0.00", it doesn't mean that it's
> absolutely zero, but rather that the load is less than 0.005.

The load caused by one call to sleep(1) and a uptime(1) every 10s is
next to nothing.  Even with the xterm displaying the valuie every 10s
(to a remote X11 server), and some daemons in the background the load
average is much less than .005.  BTW, on another server, with
Linux-2.4.37, much slower CPU (2 x Pentium III) and some more
background daemons, running emacs and gnus to write this posting, I
can run uptime every second and I get 0.00 for all 3 values (1 minute,
5 minutes, 15 minutes)

> This small load accumulates over the minutes so the 5 and 15 minute
> load averages are high enough to show.

The accumulated CPU time and other load is divided by longer time
intervals so it should still be 0.

> Try sleeping for 60 seconds between calls and see if that lowers
> things.

That doesn't change anything.  I can also wait for an arbitrary long
time, type uptime, and it will show 0.00, 0.01, 0.05 almost every
time.

I still think this is a bug.

urs


-- 
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/ygf8v4o8dod@janus.isnogud.escape.de

Wy no clang in jessie

[Urs Thuermann]

Squeeze, wheezy and sid seem to have clang packages but no jessie.  It
seems clang was removed from jessie at 2013-06-04 but I could not find
a reason.  Why doesn't jessie have a clang package?

urs


-- 
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/ygfa9mm49xc@janus.isnogud.escape.de

IP/IP and IPv6/IP tunnels and dynamic IP address

[Urs Thuermann]

My ISP provides a single, dynamic IPv4 address to me.  I have
configured several tunnels, IPv4 in IPv4 and IPv6 in IPv4, to several
sites.  This is running on a very old Linux installation and I want to
move that to a Debian system.

I know how to set up /etc/network/interfaces for such tunnels if the
IPv4 address is static.  However, with dynamic IPv4 address you do not
only needed ifup  and ifdown  but also a way to update
the remote tunnel configuration when the local IPv4 address changes.
E.g. for Hurricane Electric tunnel broker you access a certain URL
(using e.g. wget) where the URL encodes your current IPv4 address.  At
another site I need to execute a command using ssh.

I know how I can do that using cron and/or pppd configs.  But I'd like
to know if Debian has some standard way to update tunnel configs,
preferrably in /etc/network/interfaces with the rest of the interface
definition.

urs


-- 
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/ygfmwqbyfab@janus.isnogud.escape.de

Re: How do I pad files in Linux?

[Urs Thuermann]

Stan Hoeppner  writes:

> One possibility might be
> 
> ~$ cat file1 file2 >> file3
> 
> where file1 is your binary and file2 contains 64 nulls.  Now you simply
> need to create a file containing exactly 64 nulls.  I've never screwed
> with this, but I'd guess it can be done with one of the scripting
> languages, maybe a text manipulator such as awk or its variants, etc.
> Loop a print command with the proper special character syntax for null
> 64 times piping the output to file2.  You may even be able to create it
> in vi.

dd if=/dev/zero bs=1 count=64 >> file

urs


-- 
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/ygffvvyeper@janus.isnogud.escape.de

ISDN call logging

[Urs Thuermann]

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?

urs


-- 
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/ygf61pzdzhh@tehran.isnogud.escape.de

Re: ISDN call logging

[Urs Thuermann]

Zenaan Harkness  writes:

> Are you able to install your "stable" OS (Debian, Linux 2.4 kernel) on
> the newer hardware perhaps?

No, that's not an option.  The intention is to replace the old server
(Linux 2.4, not Debian) by the new server (new hardware, Debian
jessie).  Almost all services, i.e. DNS, DHCP, NTP, mail, news, and
$HOMEs + NFS, have already been migrated to the new server.  In
addition, a few VMs run on that new server, which couldn't be done
with the old Linux 2.4 system.  The only thing left on the old machine
is ISDN logging.

I also considered moving the old server to the new machine as a VM
with PCI passthrough for the ISDN card.  This also failed, AFAICT,
because the BIOS doesn't support it.

> Don't know if it's relevant, but I've played around with FreeSwitch -
> freeswitch.org

Looks interesting, but not for this problem.

urs


-- 
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/ygf7gae9blc@tehran.isnogud.escape.de

Which driver for Nvidia card on jessie

[Urs Thuermann]

After a hardware change my workstation with X11/XFCE has a new Nvidia
graphics card and the nouveau driver doesn't seem to work with new
card.  Before the hardware change I used xserver-xorg-video-nouveau.

The new Nvidia card is listed by lspci as

02:00.0 VGA compatible controller: NVIDIA Corporation Device 1430 (rev a1) 
(prog-if 00 [VGA controller])
Subsystem: Hewlett-Packard Company Device 1190
Physical Slot: 2
Flags: fast devsel, IRQ 32
Memory at f200 (32-bit, non-prefetchable) [size=16M]
Memory at e000 (64-bit, prefetchable) [size=256M]
Memory at f000 (64-bit, prefetchable) [size=32M]
I/O ports at 1000 [size=128]
Expansion ROM at f308 [disabled] [size=512K]
Capabilities: [60] Power Management version 3
Capabilities: [68] MSI: Enable- Count=1/1 Maskable- 64bit+
Capabilities: [78] Express Legacy Endpoint, MSI 00
Capabilities: [100] Virtual Channel
Capabilities: [258] L1 PM Substates
Capabilities: [128] Power Budgeting 
Capabilities: [420] Advanced Error Reporting
Capabilities: [600] Vendor Specific Information: ID=0001 Rev=1 Len=024 

Capabilities: [900] #19

I haven't found much info on which driver to use with this card.  Is
the nouveau driver able to work with it?  Or do I have to use the
Nvidia binary only driver?

urs

Re: Prevent Linux kernel from logging segfaults of user processes

[Urs Thuermann]

Steve Keller  writes:

> Since quite a long time the Linux kernel prints log messages about
> processes that cause a segmentation fault, but it has not always
> been that way (Linux 2.x, maybe also 3.x versions, AFAIR).
> 
> On a multi-user system, where users develop & debug software it is
> annoying for the admin to see all these messages in dmesg.
> 
> So my question is can these messages been turned off, either completely
> or for certain (effective or real) user ids, e.g. uid above 1000?

I think you're looking for

echo 0 > /proc/sys/debug/exception-trace

I'm not aware of a way to disable this only for certain uids.

urs

Hot swapping failed disk /dev/sda in RAID 1 array

[Urs Thuermann]

In my RAID 1 array /dev/md0 consisting of two SATA drives /dev/sda1
and /dev/sdb1 the first drive /dev/sda has failed.  I have called
mdadm --fail and mdadm --remove on that drive and then pulled the
cables and removed the drive.  The RAID array continues to work fine
but in degraded mode.

I have some questions:

1. The block device nodes /dev/sda and /dev/sda1 still exist and the
   partitions are still listed in /proc/partitions.

   That causes I/O errors when running LVM tools or fdisk -l or other
   tools that try to access/scan all block devices.

   Shouldn't the device nodes and entries in /proc/partitions
   disappear when the drive is pulled?  Or does the BIOS or the SATA
   controller have to support this?

2. Can I hotplug the new drive and rebuild the RAID array?  Since
   removal of the old drive seems not to be detected I wonder if the
   new drive will be detected correctly.  Will the kernel continue
   with the old drive's size and partitioning, as is still found in
   /proc/partitions?  Will a call

blockdev --rereadpt /dev/sda

   help?

3. Alternativley, I could reboot the system.  I have called

grub-install /dev/sdb

   and hope this suffices to make the system bootable again.
   Would that be safer?

Any other suggestions?


urs

Re: Hot swapping failed disk /dev/sda in RAID 1 array

[Urs Thuermann]

Peter Ludikovsky  writes:

> Ad 1: Yes, the SATA controller has to support Hot-Swap. You _can_ remove
> the device nodes by running
> # echo 1 > /sys/block//device/delete

Thanks, I have now my RAID array fully working again.  This is what I
have done:

1. Like you suggested above I deleted the drive (/dev/sda* and entries
   in /proc/partitions)

echo 1 > /sys/block/sda/device/delete

2. Hotplug-added the new drive.  Obviously, my controller doesn't
   support or isn't configured to notify the kernel.  Using Google I
   found the command the have the kernel rescan for drives:

echo "- - -" > /sys/class/scsi_host/host0/scan

3. The rest is straight-forward:

fdisk /dev/sda  [Add partition /dev/sda1 with type 0xfd]
mdadm /dev/md0 --add /dev/sda1
update-grub

Now, everything is up again and both drives synced, without reboot:

# cat /proc/mdstat 
Personalities : [raid1] 
md0 : active raid1 sda1[2] sdb1[1]
  1953381376 blocks super 1.2 [2/2] [UU]
  bitmap: 1/15 pages [4KB], 65536KB chunk

unused devices: 
# uptime
 11:49:01 up 106 days, 22:44, 23 users,  load average: 0.13, 0.19, 0.15

I only wonder if it's normal that the drives are numbered 2 and 1
instead of 0 and 1.

> Ad 2: Depends on the controller, see 1. It might recognize the new
> drive, or not. It might see the correct device, or not.

Next time I reboot the machine I will check whether there are any BIOS
settings to make the controller support hot-plugging.

urs

Upgrade from jessie to strech wants to bloat by system

[Urs Thuermann]

After fully updating my jessie system using

aptitude update; aptitude full-upgrade

I edited sources.list to dist-upgrade to strech.  A folloing aptitude
upgrade wants to install additional 1.5 GB on my system which is
currently ~5 GB, i.e. a 30% increase:

  # df -h / /usr
  Filesystem  Size  Used Avail Use% Mounted on
  /dev/dm-0   2.0G  1.8G  135M  93% /
  /dev/dm-2   7.8G  4.4G  3.0G  60% /usr
  # du -hs /root
  1.4G/root
  # aptitude upgrade
  Resolving dependencies...
  ...
  The following NEW packages will be installed:
btrfs-progs{a} clang-3.8{a} cpp-6{a} cpp-6-doc{a} dh-autoreconf{a}
...
  The following packages will be REMOVED:
cpp-4.9-doc{u} docutils-common{u} docutils-doc{u} g++-4.9{u} 
gir1.2-vte-2.90{u}
...
  The following packages will be upgraded:
acct acl acpi acpi-support-base acpid adduser adwaita-icon-theme apache2
...
  1151 packages upgraded, 297 newly installed, 128 to remove and 82 not 
upgraded.
  Need to get 1843 MB of archives. After unpacking 1537 MB will be used.
  Do you want to continue? [Y/n/?]

I see that some new versions of packages are installed without the old
versions being removed, although they are marked as automatically
installed, e.g. Linux kernel, clang, llvm, and some others.  For
example

  # aptitude search "~i clang"
  i   clang - C, C++ and Objective-C compiler (LLVM based)
  i A clang-3.5 - C, C++ and Objective-C compiler (LLVM based)
  i A libclang-common-3.5-dev   - clang library - Common development package
  i A libclang1-3.5 - C interface to the clang library

and aptitude full-upgrade will install clang-3.8 but not remove
clang-3.5.  But my suspicion is that even when I manually remove all
these old packages, the installation is still unreasonably larger than
it is currently.

urs

Re: Upgrade from jessie to strech wants to bloat by system

[Urs Thuermann]

Dejan Jocic  writes:

> Can you check what happens if you use apt-get update && apt-get upgrade
> followed by apt-get dist-upgrade? Aptitude is not really best tool for
> upgrades between distribution releases.

I changed to aptitude quite some time ago, because it seems to better
remove unused automatically installed packages, although it still does
not work quite satisfactory.

OK, I hope I find some time tonight, to backup /, /usr, and /var and
then do a dist-upgrade with apt-get.  But it seems not be very
different:

# apt-get --auto-remove dist-upgrade
...
1215 upgraded, 412 newly installed, 161 to remove and 0 not upgraded.
Need to get 1930 MB of archives.
After this operation, 1560 MB of additional disk space will be used.
E: You don't have enough free space in /var/cache/apt/archives/.

This command also installs e.g. clang-3.8 but does not remove the
automatically installed clang-3.5.

urs

Removing old python packages installed with pip

[Urs Thuermann]

On a machine running Debian stretch I have installed python3, which is
currently python3.5.  Nothing of python3.4 is present.

But in /usr/local/lib/python3.4/dist-packages/ a number of packages is
still installed.  Probably, these have been installed using pip3 when
python3.4 was current.

Now, it seems pip3 isn't able to remove packages from that old
directory.  Is it safe to just rm -r /usr/local/lib/python3.4?

urs

Re: Removing old python packages installed with pip

[Urs Thuermann]

Eike Lantzsch  writes:

> On Friday, December 1, 2017 3:24:47 PM -03 Urs Thuermann wrote:
> > On a machine running Debian stretch I have installed python3, which is
> > currently python3.5.  Nothing of python3.4 is present.
> > 
> > But in /usr/local/lib/python3.4/dist-packages/ a number of packages is
> > still installed.  Probably, these have been installed using pip3 when
> > python3.4 was current.
> > 
> > Now, it seems pip3 isn't able to remove packages from that old
> > directory.  Is it safe to just rm -r /usr/local/lib/python3.4?

> If those packages were installed together with the Debian system then 
> deinstall with aptitude or apt remove.

No, as I said, it was (most probably) installed using pip3 when
python3.4 was current in Debian.  Debian packagement doesn't know
about it, so cannot uninstall it.

> If you mix two different package installers the consequence is that one does 
> not know about the other and they can interfere with each other.

No, nothing is mixed.  pip installs in /usr/local, Debian packagement
installs in /usr.

But pip3 formerly installed in /usr/local/lib/python3.4 but now seems
not to know about anymore.  Now it installs in /usr/local/lib/python3.5.
Therefore, pip3 doesn't list and cannot uninstall python packages from
the older /usr/local/lib/python3.4.  Also, python3, which is 3.5
currently, seems not to search that old directory.

I am almost sure that just removing it will be fine, but I wanted to
make sure, that no file of pip3 e.g. in /var or elsewhere still refers
to /usr/local/lib/python3.4.  That's why I asked here.

> On Debian always install packages the "Debian way".

There are Python packages like theano and keras that are not in
Debian.  They can only be installed completely manually or using
pip3.  In both cases they go to /usr/local, of course.

urs

Re: Removing old python packages installed with pip

[Urs Thuermann]

deloptes  writes:

> yes it is more or less safe - it depends how you let it install - perhaps
> there is also something under /usr/local/bin

Thanks, I'll do it then.  And yes, there are 3 binaries in
/usr/local/bin from the theano package with the same installation
date, which I will also remove.

urs

Embarrassing security bug in systemd

[Urs Thuermann]

Yesterday, my 10 years old son logged into my laptop running Debian
jessie using his account, and curiously asked if he is allowed to try
the /sbin/reboot command.  Knowing I have a Linux system as opposed to
some crappy Win machine, I replied "sure, go ahead and try".  Seconds
later I was completely shocked when the machine actually rebooted...

Of course, my son doesn't have any special privileges, no entry in
/etc/sudoers, etc.  But then I see

$ ls -l /sbin/reboot
lrwxrwxrwx 1 root root 14 Apr  8  2017 /sbin/reboot -> /bin/systemctl
$ ls -l /bin/systemctl
-rwxr-xr-x 1 root root 538904 Apr  8  2017 /bin/systemctl
$ dpkg -S /bin/systemctl
systemd: /bin/systemctl

The /bin/systemctl binary is not suid root, so I assume[1] it
communicates to systemd which then reboots the machine without
checking what user the request comes from.

I wonder how can such a severe bug make it into a Debian stable
distribution?  And is this just an insane default setting on Debian's
side or is it yet another instance of brain-dead systemd behavior?

Searching the man pages I couldn't find a way to fix this.  How can
that be stopped?

[1] Of course, this is not docuemented in systemctl(1) as usual with
systemd.  Also, according to the man page, systemctl must be
called with a "COMMAND" argument which /sbin/reboot doesn't do.
Obviously, systemctl looks at the name it was called and somehow
uses that as command.  The admin shall guess about this.


urs

Hard drives with 4 KB sectors and small file systems

[Urs Thuermann]

Has anyone experience or can provide a link to information on
performance impact with the new hard drives with 4 KB sectors when
using file system with 1 KB block size?

My question is not about the alignment issue caused by
physical/logical sector size of 4096/512 Bytes.  I haven't yet played
with that but I also wouldn't expect any problems here since 4 KB
alignment should be easy to achieve.

But I have a number of ext3 and ext4 file systems that use a small
block size of 1 KB.  This is because of the average small file size on
these file systems, e.g. the news spool with an average file size of
2900 Bytes.  Going to 4 KB block size would cause an increase of the
internal fragmentation from about 15% to approx. 42% which I wouldn't
like.

But with a 4 KB sector size and 1 KB file system block size writing of
a file might decrease performance significantly.  When writing a file,
e.g. with dd if=/dev/zero of=foo bs=2900 count=1, instead of writing 3
blocks of 1 KB, i.e. 6 sectors of 512 B, it would be necessary to read
a 4 KB sector, modify 3 KB of it and write it back (assuming the 3 KB
are in the same 4 KB sector, otherwise two reads and 2 writes would be
necessary).

Has anyone practical performance figures for such a scenario?

urs


-- 
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/ygftyh7uopx@janus.isnogud.escape.de

How can I select the locale for a Gnome seesion?

[Urs Thuermann]

When I installed Debian on my laptop I chose US English as language
and selected a couple of locales to generate.  When I login using
gdm3, everything is English.  But my wife prefers German.  I know I
can run programs from the shell like LANG=de_DE foo but how can I
change the language for a whole Gnome session?

urs


-- 
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/ygfliwqsz12@janus.isnogud.escape.de

How to make X11 server listen to TCP connections?

[Urs Thuermann]

I use Gnome and gdm3 for login.  I want my X11 server to listen for
TCP connections on port 6000 from remote machines.  But I haven't
found anything in the documentation of gdm3 and in gdm's config files.

urs


-- 
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/ygfpqm3rkog@janus.isnogud.escape.de

audio CDs and cdrdao vs. cdparanoia

[Urs Thuermann]

When I rip audio CDs, I typically use both cdrdao and cdparanoia and
compare the results to make sure that I really really have the correct
digital audio data.  I run Debian testing with current versions of
cdrdao 1.2.2 and cdparanoia III release 10.2.

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.

Then I use my own small program to split the data.cdr file into wav
files .wav according to the toc file.  I then compare these wav
files with the track.cdda.wav files from cdparanoia.
Alternatively, one could run

sox track.cdda.wav cdda.cdr

and then compare data.cdr to cdda.cdr.  I most cases the results of
cdrdao and cdparanoia are the same but for roughly 1 of 4 CDs one or
more tracks differ.  Sometimes this is the case for CDs with scratches
but sometimes also for CDs with no obvious scratches where both,
cdrdao and cdparanoia don't give any error message and do not seem to
have any problems ripping the CD.  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.  From the output to stdout I see that
cdrdao uses the Paranoia DAE library and Joerg Schilling's SCSI
library to actually read the audio CDs.  With ldd I see it is not
linked dynamically to these libraries.  So one question is, do cdrdao
and cdparanoia use different library versions?

Regards,
urs


-- 
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/ygfy6dpslq7@janus.isnogud.escape.de

several debian installer questions

[Urs Thuermann]

I have a couple of questions concerning the debian installer (from the
testing distribution):

1. When downloading the iso image which I combine with boot.img.gz to
   get an image for an USB stick, I have to choose the Debian release
   (i.e. stable or testing) and the architecture I want to install on.

   I would like if I could use the same installer to install different
   releases and for different architectures.  I.e. I'd like to be to
   choose between stable and testing and between i386 and amd64 using
   the same installer instead of having 4 different installers on 4
   separate USB sticks.  Is that possible or planned?

2. Is it possible to install the debian installer onto a Debian live
   system, i.e. call the installer from a running live system?

The goal would be to have a single USB stick, which I can use to

* run a Debian live system
* to rescue systems (already possible from a customized live system)
* install Debian on Intel/AMD PCs, stable and testing, 32 bit and 64
  bit.

urs


-- 
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/ygf39qigbp1@janus.isnogud.escape.de

xlock and logout button

[Urs Thuermann]

since my last aptitude upgrade, xlock shows a logout button 10 minutes
after having locked my Xserver.  I think this is a new feature since I
don't remember having seen this behavior before.

I haven't found a way to disable the appearance of the logout button,
neither in the man page nor in /etc/X11/app-defaults.  How can I
disable this button?  I need this for a non-public terminal in my
bureau where I often lock the screen for longer than 10 minutes, going
to lunch or even over night.

urs


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

Re: Disk Defragmentatio

[Urs Thuermann]

> I know that, but I'm running a pair of debian etch boxes over vmware,
> and the disk images have grown to almost tripple the size of the
> actual installation. I want to reclaim the 3+ GBs on the partition
> (I'm running a bit short on disk space), and in order for the vmware
> utility to compact the image, I need to defrag the filesystem first.

I don't know what vmware does to provide a file system to your debian
guest.  But I'm pretty sure there is no need to defragment your file
system.  What type of file system do you use?  ext2/3, reiserfs, xfs,
...?

>From what you write, I think you want to resize your file system in
order to be able to free part of the underlying file/block device in
the host.  I case you have ext2 or ext3 you need to fsck the FS first,
then resize2fs(8).  resize2fs is part of the e2fsprogs package.  The
other file systems have similar tools.


urs


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

Re: tzfile Modifications for the United States

[Urs Thuermann]

Martin McCormick <[EMAIL PROTECTED]> writes:

>   Since the new energy bill was signed in to law, the existing
> tzfile information for all those parts of the United States that use
> DST needs to be tweaked so the automatic change will occur at the
> correct times.  Will there be new tzfile sets we can download that
> have been modified or do we need to generate our own?  The tzfile for
> any given time zone is a binary file containing constants that drive
> all the conversion routines for the timeofday functions.  It occurred
> to me that now is as good a time as any to start patching since the
> new data set won't hurt a thing to install and have running.

The source of the zone files is in the glibc sources.  There are files
in a syntax which is compiled by zic(8), the zone info compiler.  You
can change the timezone description and compile yourself, using zic,
then install the results in /usr/share/zoneinfo.


urs


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

position independent executables (PIE) on Debian?

[Urs Thuermann]

Newer versions of gcc and binutils support generation of position
independent executables (PIE) using the option -fpie for gcc and -pie
for ld.  the result is a ELF shared object that can be executed
instead of a ELF executable.  The difference is that PIE executables
can be loaded at any address and support address randomization in
order to make vulnerability exploits harder.

In Debian you can create and execute such PIE executables but their
address spaces are not randomized (I've checked with printf() and cat
/proc//maps).  I assume you need kernel support for this and the
standard Debian kernel doesn't have it.

So is there a way to get PIE executables with random addresses on
Debian?


urs


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

Re: usrmerge in bookworm

[Urs Thuermann]

Marco  writes:

> Am 06.10.2023 schrieb Steve Keller :
> 
> > I have always been sceptical about /usr merge, since all binaries now
> > appear in two places, "type sh" in bash gives the strange looking
> > /usr/bin/sh where all Uni*ers are strongly used to /bin/sh.  But also
> > things like the following don't work anymore:
> 
> Historic reasons, they don't exist anymore on current systems.
> 
> > $ dpkg -S $(type -p sh)
> > dpkg-query: no path found matching pattern /usr/bin/sh
> 
> That is because /bin is now a symlink to /usr/bin and the file isn't
> provided in /usr/bin by the package.

I assume Steve (seemingly multi Un*x user) is aware about the historic
reasons and how and why /usr-merge is implemented using symlinks.

> > And I don't see a comfortable way around this.
> 
> One way would be to strip off /usr from the string to search.

That, of course, doesn't really help

$ type -Pa ls
/usr/bin/ls
/bin/ls
$ dpkg -S $(PATH=/bin type -Pa ls)
coreutils: /bin/ls
$ type -Pa gcc
/usr/bin/gcc
/bin/gcc
$ dpkg -S $(PATH=/bin type -Pa gcc)
dpkg-query: no path found matching pattern /bin/gcc

because some packages have their binaries in /bin and others in
/usr/bin, and likewise with /sbin vs. /usr/sbin, and /lib
vs. /usr/lib.

But the -a option to bash's 'type' builtin command helps:

$ dpkg -S $(type -Pa ls gcc)
dpkg-query: no path found matching pattern /usr/bin/ls
coreutils: /bin/ls
gcc: /usr/bin/gcc
dpkg-query: no path found matching pattern /bin/gcc

urs

Re: usrmerge in bookworm

[Urs Thuermann]

Greg Wooledge  writes:

> Yeah, usrmerge is a bit wonky in these early stages.  

$ apt-get changelog usrmerge | tail -n2
 -- Marco d'Itri   Tue, 04 Nov 2014 22:42:44 +0100
Fetched 11.0 kB in 0s (58.9 kB/s)

Not what I'd call 'early' stages.

> Part of the reason for this is that it's not *mandatory*, not
> really.

That seems to have changed in Debian 12:

$ aptitude show init-system-helpers | egrep ^Prio\|^Dep
Priority: required
Depends: usrmerge | usr-is-merged

In Debian 11, init-system-helpers did not depend on usrmerge.

> There are still Debian systems -- important ones, which run the
> internal infrastructure of Debian itself -- that aren't using it.

Are those running older versions of Debian than stable?

urs

Re: Strip domain name from hostnames in output

[Urs Thuermann]

Michael Stone  writes:

> On Mon, Sep 16, 2024 at 11:18:19AM +0200, Steve Keller wrote:
> >Is there a config option in libc, the host name resolver or somewhere
> >else to show hostnames in my own domain without the full domain name?
> 
> no

You could at least achieve that desired behavior by putting all your
local hosts to /etc/hosts without domain name.  You would have to do
this on all your hosts (or NIS aka YP, if you have that).  Ugly, if
you already have DNS running.

The getnameinfo(3) function has a flag NI_NOFQDN which would do what
you want for local hosts, but that doesn't help here.

So, missing is only a global config flag for getnameinfo() that let's
it behave as if NI_NOFQDN had been specified if the domain name
matches the one in /etc/resolv.conf.  Useful would also be another
flag to strip domain names that are listed in /etc/resolv.conf in the
"search" entry.  I also would like such a functionality.

urs