I am not quite sure what you mean, but in this case systemd is *not* init.
elogind conflicts with systemd.
Regards
Harri
PS: Workaround is
chmod -x /etc/init.d/cgroupfs-mount
start over and reboot.
Hope this helps
Harri
Package: cgroupfs-mount
Version: 1.4
Trying to uninstall lxc and some related packages I got an error from
cgroupfs-mount:
:
:
Removing cgroupfs-mount (1.4) ...
Unmounting cgroupfs hierarchyumount: /sys/fs/cgroup/elogind: target is busy.
invoke-rc.d: initscript cgroupfs-mount, action "stop" fai
If Debian introduces a new feature, changes an API or something like this,
breaking POSIX compliance, is this a bug?
I grew up with several UNIXes listed on the compliance web page. I never
understood why df and others had to work different on some Linux distros. IMHO
this set of common basic
Package: debian-policy
Version: 4.4.1.2
I haven't found it mentioned in the policy manual, so I wonder if
Debian is supposed to be POSIX compliant (unless noted otherwise)?
IMHO a "it goes without saying" is not sufficient.
Is it safe to assume that POSIX code works?
Regards
Harri
This is still WIP. See https://wiki.strongswan.org/issues/3197
I would be highly interested, too. deb-multimedia.org is not an option.
Thanx in advance
Harri
metoo
Its a pretty huge effort to roll out a wrapper around freeipa to +300
Hosts and keep it maintained for 2 years until Bullseye is released.
I am using freeipa-client on Debian for several years now. This problem
never came up before, even though it was Python 2 as well, so I wonder if
there could
Package: freeipa-client
Version: 4.7.2-3
ipa complains about a locale that is not set:
% ipa user-show --all
ipa: ERROR: System encoding must be UTF-8, 'ANSI_X3.4-1968' is not supported. Set LC_ALL="C.UTF-8",
or LC_ALL="" and LC_CTYPE="C.UTF-8".
% env | grep ANSI
% env | grep 1968
% env | egrep
PS: One important aspect is that /etc/default/opensmtpd is *not* included
in the Debian package. If it would be included, then I would get conflicts
again.
Regards
Harri
Hi Ryan,
On 2019-11-23 18:07, Ryan Kavanagh wrote:
I'm trying to understand the rationale for having a separate smtpd.conf.
/etc/smtpd.conf is listed in conffiles, so package upgrades will never
clobber the changes you have made to it. During upgrades, if a package
has a new version of /etc/smt
Attached is a better version, introducing /etc/default/opensmtpd.conf.
Regards
Harri
#!/bin/sh
# /etc/init.d/opensmtpd
#
# Written by Daniel Walrond
#and Ryan Kavanagh
### BEGIN INIT INFO
# Provides: opensmtpd mail-transport-agent
# Required-Start:$local_fs $remote_fs $sys
Package: opensmtpd
Version: 6.6.1p1-1
Severity: wishlist
Would it be possible to make the opensmtpd config file configurable?
Some code in the init file like
test -f /etc/default/opensmtpd.conf && source
/etc/default/opensmtpd.conf
: ${CONFIG:="/etc/smtpd.conf"}
:
Package: clamav
Version: 0.101.4+dfsg-1
Tag: security
Upstream provides a new version 0.102.1, including a patch for
CVE-2019-15961
See https://blog.clamav.net/2019/11/clamav-01021-and-01015-patches-have.html
Regards
Harri
I cannot confirm the fix. Even when rpcsec_gss_krb5 *is* loaded,
rpc-svcgssd.service still fails for Buster:
root@srvl064:/etc# systemctl daemon-reload
root@srvl064:/etc# systemctl restart rpc-svcgssd
Job for rpc-svcgssd.service failed because the control process exited with
error code.
See "sys
Package: logrotate
Version: 3.14.0-4
Severity: wishlist
".bak" is the traditional extension for local backup files used by various
editors (e.g. mg). Do you think the *.bak files in /etc/logrotate.d could
be ignored by default, similar to *~ and others?
Of course I have found the tabooext option
I'd love to help maintaining this package. Since this bug
is a show stopper for me I am running my own opensmtpd
package anyway. Not to mention it that I am using opensmtp
on OpenBSD, too.
Regards
Harri
Package: linux-image-4.19.0-6-amd64
Version: 4.19.67-2+deb10u1
When I switched off my USB disk dockingstation (2 slots, 2 Sata
disks were inserted) I got a kernel dump:
Oct 31 16:26:44 dpcl082 kernel: [22985.418721] usb 6-2: USB disconnect, device
number 2
Oct 31 16:26:44 dpcl082 kernel: [22985
Have you tried to reinstall the xinit and xinput packages?
Regards
Harri
PS: IMHO its unreasonable to suspend the host by default, just
because gdm3 is installed. sshd is installed, too, but obviously
it has no say in this case. Why has gdm3?
I would suggest to make 'blank' the default at least for ac mode.
Regards
Harri
metoo.
Cannot say if its exactly 20 minutes, but after I moved to gdm3 the
desktop PCs running Debian 10 became unresponsive after some time, if
nobody was logged in.
This seriously affects remote administration of about 70 Linux desktop
PCs. I moved back to lightdm due to this bug.
Regards
Ha
Package: drbd-utils
Version: 9.5.0-1+b1
manpage for drbd-overview says
This tool will be deprecated soon. Please consider using drbdtop.
Would it be possible to either *keep* drbd-overview (I am using it
for monitoring) or to add drbdtop to Debian?
Thanx in advance
Harri
Stay tuned, I had a busy week.
But I am surprised that you rely on "non-Debian" software to
reproduce the bug. Ain't the test case mentioned in
https://github.com/systemd/systemd/commit/37ed15d7edaf59a1fc7c9e3552cd93a83f3814ef
sufficient to reproduce the problem and verify the fix, without
depe
Package: systemd
Version: 241-5
systemd in Buster and older fails to handle MountFlags correctly,
see
https://github.com/systemd/systemd/commit/37ed15d7edaf59a1fc7c9e3552cd93a83f3814ef
This seriously affects the "containerd" versions of Docker >= 18.09.
On installing Rancher (just as an example
The 20 char lower limit for PSKs has been dropped by upstream.
https://github.com/strongswan/strongswan/pull/145
FYI:
I had problems with linux-image-4.19.0-5-amd64, too, esp. on writing
large amounts of data to an USB block device (USB-C). "sync" got stuck,
umount didn't work, host became unresponsive, etc. Reset necessary.
linux-image-4.19.0-6-amd64 (in proposed-updates) seems to work better.
Regards
Ha
upstream provides a "portable" branch with support for
openssl 1.1.1c on github.
Regards
Harri
Package: apt-xapian-index
Version: 0.49
update-apt-xapian-index dies with "index out of range":
root@git01:~# dpkg -P apt-xapian-index
(Reading database ... 58218 files and directories currently installed.)
Purging configuration files for apt-xapian-index (0.49) ...
Removing index /var/lib/apt-
On 8/10/19 12:52 PM, Sven Joachim wrote:
> Am 10.08.2019 um 10:34 schrieb Harald Dunkel:
>
>> Package: aptitude
>> Version: 0.8.11-7
>> Severity: wishlist
>>
>> Probably its difficult to implement, but what I am *really*
>> missing in aptitude is a lis
On 8/10/19 2:37 PM, Axel Beckert wrote:
> If I recall correctly, the abandoned 0.6.9 branch had a feature where
> you can select the "Source Package" line in a binary package view and
> then press Enter on it ("[" would be another, probably more helpful
> option) and then you got a list of all bina
Package: aptitude
Version: 0.8.11-7
Severity: wishlist
Probably its difficult to implement, but what I am *really*
missing in aptitude is a list of source packages. If I press
'[' it should expand to list all its binary packages to
chose from.
This could help alot especially on large packages (e.
Hi Rupert,
I understand that this is highly frustrating. Apparently you
have found the hidden radio buttons for the "Password" entry?
Regards
Harri
Package: strongswan
Version: 5.7.2-1
Upstream provides a new strongswan version 5.8.0, including
IPv6 support for charon-nm. See
https://wiki.strongswan.org/issues/1143
Would it be possible to include this version for Bullseye?
Thanx in advance
Harri
Package: ieee-data
Version: 20150531.1~deb8u2
The monthly cron job returns
# export BASEDIR=/var/lib/ieee-data/
# bash -x /usr/bin/update-oui -f -q
:
:
+ goAndGet http://standards.ieee.org/regauth/oui/oui.txt oui.txt
+ true
+ wget -q -O- http://standards.ieee.org/regauth/oui/oui.txt
+ Die 'wget
Package: dpkg
Version: 1.18.25
Severity: wishlist
On a package upgrade/reinstall dpkg should show a warning about
a missing config file that is not recovered from the *.deb file.
It is pretty unlikely that a config file has been deleted on
purpose. It is much more likely that a config file has b
Obviously the list of processes (as produced by ps -ef, pgrep, whatever)
does not provide sufficient information to find the right process id to
send a signal to. Additional information has to be included.
AFAICS thats the PID file in /run. If this file doesn't exist, then
the only reasonable ass
Any news about this? How comes this wasn't an issue for Jessie?
Regards
Harri
Package: dpkg
Version: 1.18.25
On the upgrade from Stretch to Buster there was a config
file conflict (with /etc/issue, but that doesn't matter).
dpkg allowed me to examine the conflict ("Z"). I could
adjust the old file using the mg editor, I could save the
config file using Ctrl-X Ctrl-S, but I
PS: AFAIR a second SIGHUP to the same process might get lost, if it
was sent shortly after the first one. Since rsyslog is responsible
for several log files, this might be the case here.
Regards
Harri
On 7/3/19 6:05 PM, Michael Biebl wrote:
> Am 03.07.19 um 17:45 schrieb Harald Dunkel:
>> I have commented out the delaycompress in all logrotate config files (>10)
>> on all my Debian hosts (>300) running squeeze and newer.
>>
>> No problem since then, AFAICT.
>
I have commented out the delaycompress in all logrotate config files (>10)
on all my Debian hosts (>300) running squeeze and newer.
No problem since then, AFAICT.
Regards
Harri
Its sad that Debian missed the train.
Regards
Harri
Package: dpkg
Version: 1.19.7
Obviously
export DPKG_PAGER=""
doesn't work as expected (overriding $PAGER, as promised by the man page).
OTOH Stretch and Jessie give me
% dpkg --no-pager -l
dpkg: error: unknown option --no-pager
Type dpkg --help for help about i
How comes that this patch hasn't made it into Buster?
Got trapped in the systemd vs no-systemd debate instead
of providing a good package?
Harri
No problems by now. Currently I am using version 5.1.7, built from upstream's
sources.
Kernel 5.0 has reached EOL.
Regards
Harri
Package: devscripts
Version: 2.19.5
Sometimes I have to archive source and binary packages, or to pickup
packages from backports to add them to a local repository. Do you
think it would be possible to add some kind of "mirror" mode to dget
to download all, any, ${arch} binary packages and the sour
I tried it (using sid and upstream's kernel 5.1.6): Builds fine,
X appears to work.
Regards
Harri
AFAICS its documented in start-stop-daemon(8) for "start-stop-daemon -x":
"Take into account processes running from inside a chroot will also be
matched, so other match restrictions might be needed."
Regards
Harri
Apparently its already documented, so please excuse the noise.
Harri
Package: dpkg
Version: 1.18.25
AFAICS a line
start-stop-daemon -x /usr/sbin/sshd -T
returns 0 (success), if it finds the sshd running in a container,
even if the local service has been stopped. Sample session (running
2 LXCs with their own private sshd each):
root@srvl011a:~# ps -ef |
Works very well.
But how comes that this report has been marked as fixed? Its not
fixed, it has just been moved out of sight to experimental to avoid
a decision. :-(
Sorry to say, but other distros are better in fixing bugs.
Regards
Harri
Are other Intel NICs affected by this problem as well?
Regards
Harri
metoo
It would be very nice if this could be fixed.
metoo
Using upstream's google-chrome 74.0.3729.108 there is no such
problem.
I haven't seen this problem for kodi 2:17.6+git20180430.1623-final-00.1
all. Apparently it has been resolved.
Package: jaxws
Version: 2.3.0.2-1
Trying to backport jaxws to Stretch I stumbled over this:
{hdunkel@dpcl082:jaxws-2.3.0.2 () 522} fakeroot debian/rules clean
dh clean --buildsystem=maven
debian/rules override_dh_auto_clean
make[1]: Entering directory
'/export/local/debian-backports/stretch/j
Hi Colin,
sorry for the delay.
On 3/23/19 2:57 PM, Colin Watson wrote:
>
> I think you should find that a good workaround is to mount /run in the
> chroot. This is normally a good idea anyway if you're doing anything
> complicated in the chroot.
>
I am not sure if this is a good idea. /run co
The patches for 6.4 might still need some improvements: I tried
opensmtpd on Unstable today, but it died with
Apr 11 14:29:55 sylvester smtpd[21692]: pony express: unxepected call to
smtp_sni_callback()
on the first EMail to relay.
Regards
Harri
lvm2 is version 2.03.02-2
PS: grub in the chroot is version 2.02+dfsg1-13 .
I apologize again for not sending the log file you asked for right
from the start. Hopefully you found the second strace log file?
Of course it is vgs that got slow somehow, but AFAICS it is called
about 16 times for each grub-mkconfig, even though there are only 4
logical volumes. Grub-mkconfig w
PPS: There is a lot of screen output of the upgrade session:
:
:
Setting up mdadm (4.1-1) ...
update-initramfs: deferring update (trigger activated)
Generating grub configuration file ...
File descriptor 3 (pipe:[145329]) leaked on vgs invocation. Parent PID 31909:
/usr/sbin/grub-probe
File desc
PS: strace reveals:
:
access("/sys/devices/pci:00/:00:1d.2/:40:00.0/nvme/nvme0/nvme0n1/uevent",
F_OK) = 0
readlinkat(AT_FDCWD,
"/sys/devices/pci:00/:00:1d.2/:40:00.0/nvme/nvme0/nvme0n1/subsystem",
"../../../../../../../class/block", 4096) = 32
openat(AT_FDCWD,
"/sys/de
Package: grub-common
Version: 2.02+dfsg1-13
Running an upgrade from Stretch to Testing in a chroot grub-probe
appears to be stuck:
root 6373 4864 0 12:56 ?00:00:04 \_ sshd: root@notty
root 6382 6373 0 12:56 ?00:00:05 \_ xterm
root 6385 6382 0 12:56 pts/
Package: monitorix
Version: 3.10.1-1
Severity: wishlist
Monitorix providing its own web service is surely a major
improvement, but using port 8080/tcp per default has a high
chance to introduce a conflict with other tools, esp. on
test systems run by unprivileged users. Port 8080/tcp is
the first
I am running ext4 instead of reiserfs today, but logging fsck has still
a *severe* impact on boot time performance. We have a few Debian file
servers in the office, e.g. providing /home/* via NFS. They are managed
remotely using some serial-over-line technology instead of a vga console.
A few mont
Package: bind9
Version: 1:9.10.3.dfsg.P4-12.3+deb9u4
"/etc/init.d/bind9 stop" kills a named running in a LXC
container, if there is no named running on the host. Same
for "restart". Sample session (on the host):
# ps -ef | grep named
opensmt+ 2349633421 0 11:14 ?00:00:00 /usr/sbin/
Hi Simon,
On 3/11/19 10:04 AM, Simon McVittie wrote:
I assume from the version that you're using stretch? Please report
bugs with reportbug whenever possible, that way I'd already have this
information.
ACK
Do you have a line for pam_limits.so in /etc/pam.d/systemd-user?
Yes.
Do you
Hi Pierre,
On 3/8/19 2:30 PM, Pierre-Elliott Bécue wrote:
I'm sorry, but lxc-templates is not needed to have lxc3 working, so it
won't become a Dependency. Moving the files from src:lxc-templates to
src:lxc and subsequently creating two DS uploads is not a fine solution
either.
lxc-templates
Package: lxc
Version: 1:3.1.0+really3.0.3-5
Running
lxc-create -t download -- --help
I had expected to get usage information about the download template,
but instead I got a huge list of images. Next it showed
Distribution:
and got unresponsive. Exit with ^C or ^D. Apparently it
Hi Antonio,
wven if Recommends is turned off, its obvious that the template
specific include files are *required* for a successful migration
to lxc 3. It shouldn't matter that Recommends is turned off.
Platform specific template scripts might be considered as deprecated
by upstream [1], but
* t
Package: lxc
Version: 1:3.1.0+really3.0.3-5
After the upgrade to lxc 3 the old containers created by
lxc 2 don't work anymore, even though lxc-update-config has
been run (manually). The config files still contain template
specific include statements, e.g.
lxc.include = /usr/share/lxc/con
Sorry to reopen this bug report. The issue is still unresolved.
And its not a documentation issue, either.
Its not reasonable to install and activate an ssh service at
debootstrap time, and then disable it for the only available
login because ssh password-based login is vulnerable to a brute
forc
Sorry, this sounded very harsh. What I meant was:
dbus-user-session sneaked in by some package recommendation
recently, AFAICT. It should not affect a working configuration
in /etc/security/limits.d/ created back in 2016.
Regards
Harri
Apparently this comes up only, if dbus-user-session (1.10.26-0+deb9u1)
is involved. I did not ask for this package. After deleting it
the problem was gone.
Regards
Harri
Package: gnome-terminal
Version: 3.22.2-1
My /etc/security/limits.d/local.conf says
* soft nofile 4096
* hard nofile 16584
root soft nofile 4096
root hard nofile 16584
But within gnome-terminal I just get
% egrep ^Limit\|open /proc/$$/limits
Limi
Package: login
Version: 1:4.4-4.1
Somehow /var/log/lastlog grew to a 110 TByte sparse file:
# ls -al /var/log/lastlog
-rw-rw-r-- 1 root utmp 110784817228 Feb 22 08:37 /var/log/lastlog
# du -h /var/log/lastlog
88K /var/log/lastlog
This affects rsync performan
Metoo.
Hopefully this is not too difficult to fix before Buster?
Regards
Harri
Package: logrotate
Version: 3.14.0-4
Severity: wishlist
I would suggest to set a default maxsize in /etc/logrotate.conf.
100M (or even higher) would do.
It is highly painful if you run into a full disk because some
mad daemon flooded your /var/log partition within a few minutes,
and the next rot
Hi Ryan,
of course I understand the risk, esp since there will be a new
opensmtpd version every 6 months, and Buster has to run for a
long time.
I would be glad to have a version in experimental.
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Package: linux-image-4.19.0-2-amd64
Version: 4.19.16-1
I am hit by a general protection fault in netfilter, see
https://marc.info/?l=netfilter&m=154807599219095 .
According to upstream the problem has been fixed in 4.19.17.
Regards
Harri
Package: opensmtpd
Version: 6.0.3p1-4
Tags: patch
If ROOT_ADDRESS is not set, then postinst fails in line 144
with "[: argument expected".
Patch is attached. I would suggest to rewrite all test statements
around $ROOT_ADDRESS in this way.
Regards
Harri
--- debian/postinst.bak 2018-01-16 16:16:
Package: dpkg
Version: 1.19.4
I got this error:
# apt upgrade
Reading package lists... Done
Building dependency tree
Reading state information... Done
Calculating upgrade... Done
The following packages will be upgraded:
libdbus-glib-1-2 libdbus-glib-1-dev libdbus-glib-1-dev-bin python-dbus
4 up
Package: opensmtpd
Version: 6.0.3p1-4
Upstream provides a new version OpenSMTPD 6.4.1, see
https://opensmtpd.org/announces/release-6.4.0.txt
https://opensmtpd.org/announces/release-6.4.1.txt
3 important changes to version 6.0.3:
- requires libressl
- config file syntax has been reworked, breaki
Package: dhcpcd5
Version: 7.0.8-1
Severity: Grave
After the upgrade dhcpcd5 appears to be broken:
# dpkg -i ~/dhcpcd5_7.0.8-1_amd64.deb
(Reading database ... 276261 files and directories currently installed.)
Preparing to unpack .../root/dhcpcd5_7.0.8-1_amd64.deb ...
Unpacking dhcpcd5 (7.0.8-1)
Package: lxc
Version: 1:3.0.3-1
lxc's sysvinit script says explicitly
# Required-Start: $syslog $remote_fs
# Required-Stop: $syslog $remote_fs
in the LSB header.
Apparently the lxc.service file for systemd has lost this dependency:
After=network.target lxc-net.service
I just had the case that cron ran a user's @reboot job, even
though his $HOME wasn't mounted via NFS yet.
init was systemd.
Regards
Harri
metoo. I'd love to see libressl available for Debian. Newer versions of
opensmtpd (coming from the OpenBSD world as well) dropped support for
openssl in favor of libressl, see.
https://poolp.org/posts/2018-11-03/opensmtpd-released-and-upcoming-filters-preview/
Regards
Harri
metoo
Sample (run on my laptop):
# pvdsplay
WARNING: Device /dev/sda not initialized in udev database even after waiting
1000 microseconds.
WARNING: Device /dev/vg00/root not initialized in udev database even after
waiting 1000 microseconds.
WARNING: Device /dev/sda1 not initializ
Package: virtualbox
Version: 5.2.22-dfsg-2
Do you think it would be possible to include the new major
release 6 in Buster? Is it too late already?
Regards
Harri
On 1/10/19 12:02 PM, Pierre-Elliott Bécue wrote:
Sorry for the inconvenience.
Not an inconvenience at all. I highly appreciate your
aspiration to keep lxc in Debian up-to-date.
My suggestion would be to move lxc 3.1 to experimental.
Thanx very much
Harri
I don't want to complain, but since version 3.1 in sid is *not*
LTS I am not sure if this bug has been resolved.
Is it wise to skip version 3.0 LTS for Buster?
Regards
Harri
Package: cron
Version: 3.0pl1-130
Apparently the users' @reboot cron jobs might be run before the
system init scripts have been completed. Sample (using sysv init):
# ls -al /etc/rc2.d/
total 20
drwxr-xr-x 2 root root 4096 Jan 2 14:44 .
drwxr-xr-x 153 root root 12288 Jan 2 14:44 ..
lrwxrwxr
Instead of showing a warning about the bug, cron should be fixed.
Just add a dependency to the most important directory services
(ldap, nis, whatever), making sure that cron is started *after*
the directory services.
Or simply start cron daemon last at boot time.
Regards
Harri
Hi folks,
how good are chances to get LXD into Buster?
Regards
Harri
Hi Julian,
On 12/12/18 1:56 PM, Julian Andres Klode wrote:
That's what it already does; unless it is an upgrade, then it
does not change the state (or, it only changes the autostate if there
is no other state change it can do).
Confirmed. I saw this too late.
Nevertheless, aptitude removed
Package: apt
Version: 1.4.8
I ran "apt install symlinks" on Stretch, even though it was
already installed due to a package dependency. When this package
dependency changed (the referring package was removed),
symlinks was removed automagically. :-(
This is not as expected. "apt install mypackage
PS: Its "invoke-rc.d", not "update-rc.d".
Sorry, my bad.
Regards
Harri
Hi Timo,
this "fix" breaks our existing HA systems based upon sysvinit-core.
:-(
By now I could install systemd, run ipa-client-install, replace
systemd with sysvinit-core, and continue setting up the system.
Now I've got a hard-wired dependency upon systemd (breaking our
existing HA systems), e
301 - 400 of 1866 matches
Mail list logo