Package: meld
Version: 3.20.2-2
If I have to interrupt my current task to check something on another
git repository, then the meld used for the current task is terminated
as soon as I hit ^Q in the meld started for the interrupt task.
This is confusing, not being asked for, and *highly*
Package: opensmtpd
Version: 6.8.0p2-3
I have to distinguish between root and postmaster EMails on my
servers, so the common option
Root and postmaster mail recipient:
in the installation menu for opensmtpd (and others) is not help-
ful. I have to manually fix this again and again.
Sorry, my bad. On Unix it is not uncommon to use file names like
myscript.sh or anotherscript.py. I guess I read what I wanted to
read.
Regards
Harri
Package: debianutils
Version: 4.11.2
If I modify /etc/crontab to use run-parts --verbose, then I get these
messages per EMail:
run-parts: executing /etc/cron.daily/01_hello
Thu Jun 30 06:25:01 CEST 2022
hi mom
run-parts: /etc/cron.daily/01_hello exited with
Package: diffutils
Version: 1:3.7-5
Upstream provides version 3.8. Do you think you could include it into
Debian for Bookworm?
Thank you very much in advance
Harri
Package: dpkg
Version: 1.20.9
Severity: wishlist
To avoid confusion the man page to dpkg-statoverride should
mention explicitly that it does *not* run chmod or others to
actually change the access permissions to already installed
files, or to files that are not part of a Debian package, or
to
I think you missed another case: A third package might install a
diversion for /bin/sh.
Regards
Harri
Package: kodi
Version: 2:19.4+dfsg2-1
Severity: wishlist
To reduce the footprint of kodi and since wayland is not *that* stable yet,
could you please provide a "non-wayland" kodi package as well? Setting
BUILD_WAYLAND := no
seems to be sufficient.
Thank you very much in advance
Harri
Package: x11-utils
Version: 7.7+5
Hi folks,
If I set
XTerm*locale: true
then xterm requires /usr/bin/luit, which can be found in x11-utils.
x11-utils puts appr 150 MByte additional files and directories into
/, even if Install-Recommends is set to false.
Would it be possible to move
Package: xdg-desktop-portal-gtk
Version: 1.14.0-1
I experience long startup delays of some gtk apps, e.g. handbrake-gtk.
Apparently there are problems with a portal service:
% systemctl --user status xdg-desktop-portal.service
x xdg-desktop-portal.service - Portal service
Loaded: loaded
Hi Vasyl,
using kodi version 2:19.4+dfsg1-2 I don't have an issue right now.
Thank you very much
Harri
Package: qmmp
Version: 1.4.4-1
Upstream provides new versions 1.5.4 and 2.0.4, see
https://qmmp.ylsoftware.com/
https://sourceforge.net/projects/qmmp-dev/
Would it be possible to upgrade the Debian package? I can help.
Regards
Harri
After rebuilding bzflag with libsdl2-2.0-0 version 2.0.20+dfsg-2 the
problem went away.
How can I get out of this? What do you suggest? Are the tools to
create custom locales included in the Debian package by accident?
Harri
Package: xserver-xorg-core
Version: 2:1.20.11-1
Severity: wishlist
Using clusterssh I ran into a "256 connections max" limit of the
XWindow server. The error message was
Maximum number of clients reached
xterm: Xt error: Can't open display: :4
First suspect was a bad ulimit,
see https://bugzilla.mindrot.org/show_bug.cgi?id=3411
Package: kafkacat
Version: 1.6.0-1
Upstream has changed the name and provides a new version 1.7.0:
https://github.com/edenhill/kcat
Would it be possible to update the package in Debian? I wouldn't
care if the source and binary packages are still called kafkacat.
Thank you very much
After installing the intel-media-va-driver package it worked until
yesterday. Today it doesn't, but that seems to be the fault of a new
libigdgmm12 package.
Regards
Harri
kodi is affected, too. Attached you can find kodi's log file.
AFAICT it *does* try vdpau first, but then it exits.
Regards
Harri
kodi.log.gz
Description: GNU Zip compressed data
Package: apt
Version: 2.4.1
Severity: wishlist
apt update gives me several messages like
W: http://example.com/debian/dists/unstable/InRelease: Key is stored in legacy
trusted.gpg keyring (/etc/apt/trusted.gpg), see the DEPRECATION section in
apt-key(8) for details.
W:
PS: No such problem for openssh-client 1:8.8p1-1 from Testing.
Package: openssh-client
Version: 1:8.9p1-3
If I use
ssh -XAf localhost xterm
to run a second xterm, then ssh eats up 100% CPU after closing the
first xterm.
Same happens for xeyes or other XWindow apps. It doesn't have to be
localhost, either.
"nohup ssh -XAf ..." seems to work, but
Instead of dropping this bug report with won'tfix after 8 years it would have
been appropriate to reassign it to the useradd package immediately.
Thanx very much for your help
Harri
Found it on https://forum.ubuntuusers.de/topic/va-api-meldet-fehler-bei-vainfo/:
The intel-media-va-driver package was missing, as it seems.
Thanx for your help
Harri
See attachments.
Regards
Harri
glxinfo.txt.gz
Description: GNU Zip compressed data
libva info: VA-API version 1.14.0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/iHD_drv_video.so
libva info: va_openDriver() returns -1
libva info: Trying to open
See attachments.
Regards
Harri
glxinfo.txt.gz
Description: GNU Zip compressed data
libva info: VA-API version 1.14.0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/iHD_drv_video.so
libva info: va_openDriver() returns -1
libva info: Trying to open
This is X. I have attached the logfile. Hope this helps.
Regards
Harri[ 3.948]
X.Org X Server 1.21.1.3
X Protocol Version 11, Revision 0
[ 3.948] Current Operating System: Linux lola.afaics.de 5.16.10-raw #1 SMP
PREEMPT Wed Feb 16 18:04:47 CET 2022 x86_64
[ 3.949] Kernel command
Package: kodi
Version: 2:19.3+dfsg1-1+b1
kodi doesn't show on my shiny new Intel NUC NUC11TNKi3. No splash screen.
No input menu. No crash AFAICS, it just exits. I don't even have a chance
to modify kodi's settings.
Logfile is attached.
Regards
Harri2022-03-01 18:21:31.433 T:1040 INFO :
Package: grub-efi-amd64
Version: 2.04-20
Trying to migrate from MBR to uefi I prepared an ESP (vfat, "esp" tag
is set, mounted on /boot/efi) and replaced the grub-pc and grub-pc-bin
packages by grub-efi and grub-efi-amd64.
Problem: The postinst script of grub-efi-amd64 seems to update
Package: handbrake
Version: 1.5.1+ds1-1
I have installed all packages from the Recommends list (the Suggests
list is empty), I am using handbrake in a local Xwindow session, dbus
and systemd are installed, and yet handbrake cannot shutdown the host
after completing the queue. Last words: A popup
Package: etckeeper
Version: 1.18.16-1
Severity: wishlist
I understand that I can tell etckeeper to not checkin changes before
installing and upgrading packages, but it might be helpful to avoid an
automatic checkin *after* package upgrade somehow, e.g. to handle a
conflict in a config file using
Hi Fabio,
sorry, but this BR is >10 years old. I don't have the hardware anymore.
Regards
Harri
I would say this is fine now (for systemd 247.3-6 in Stable).
Regards
Harri
Package: systemd
Version: 247.3-6
I have used debootstrap to install Debian on an old qnap. Problem:
The power button appears to be ignored, i.e. I have to login remotely
to run poweroff. Too bad that PermitRootLogin in sshd_config doesn't
allow.
This is basic functionality and should work
Package: lxc
Version: 4.0.11-1~exp4
Did you notice version 4.0.12 on https://linuxcontainers.org/lxc/downloads/ ?
See https://discuss.linuxcontainers.org/t/lxc-4-0-12-has-been-released/13288
Regards
Harri
Sure, but why does systemctl complain about rpc-svcgssd.service, if
there are no NFS credentials inside? Its perfectly fine to run NFS
without cryptographic security, even though /etc/krb5.keytab does
exist.
I have the impression that the conditionals you mentioned are not
complete.
Package: lxc
Version: 1:4.0.10-1
Due to some RC bugs in bridge-utils lxc and lxc-templates might be dropped
from Testing and Sid. I would suggest to drop this dependency instead, in
favor of "ip bridge". According to
https://wiki.linuxfoundation.org/networking/bridge
the bridge-utils package is
upstream bug report: https://github.com/BZFlag-Dev/bzflag/issues/301
Maybe the Recommends could be replaced by a Suggests: x11-utils.
By now I didn't even know that there is a tool "luit".
Apparently this has bee triggered by the upgrade of libsd2-2.0-0. If I downgrade
this package to the version in stable, the bzflag works again.
```
Aptitude 0.8.13: log report
Sun, Jan 23 2022 17:40:07 +0100
IMPORTANT: this log only lists intended actions; actions which fail
due to dpkg
Package: bzflag
Version: 2.4.22-1
Apparently bzflag gets stuck on Sid, eating up 80% CPU on one core. I can
move the mouse, but keyboard input is ignored, even F12. Rebuilding the
package within Sid did not help.
There is no such problem with bzflag on MacOS, i.e. the network connection
seems
Package: certbot
Version: 1.21.0-1
Upstream provides a new version 1.22, see
https://github.com/certbot/certbot/releases.
Would it be possible to support the new version for Sid?
Thank you very much
Harri
Package: nfs-kernel-server
Version: 1:1.3.4-6
I would like to reduce nfsv4gracetime and nfsv4leasetime from 90 sec
to something reasonable. Looking through several scripts and systemd
config files I figured out that I have to set
RPCNFSDOPTS
in /etc/default/nfs-server.conf to make a
Package: nfs-common
Version: 1:1.3.4-6
systemd moans about krb5.keytab at boot time
```
# systemctl --failed
UNITLOAD ACTIVE SUBDESCRIPTION
* rpc-svcgssd.service loaded failed failed RPC security service for NFS server
LOAD = Reflects whether the unit definition was
Control: tag -moreinfo
The caches have to be invalidated before nscd is removed, because
after removing nscd (the binary) you cannot run "nscd -i" anymore.
The cache files have to be removed to make sure that new jobs ignore
the caches completely.
A general problem with nscd is that running
On 2021-11-19 19:18:37, Jeremy Sowden wrote:
Some of those seg-faults occurred in libnftnl.so.11.2.0, the remainder
in libnftnl.so.11.3.0. Are all those log messages from one node? The
libnftnl11 package in Debian 11 (1.1.9-1) contains libnftnl.so.11.5.0.
What have you got installed?
Package: iptables
Version: 1.8.7-1
iptables-nft gives me SIGSEGVs on the nodes of my Kubernetes
cluster, e.g.
:
[Fri Nov 12 11:43:46 2021] iptables-save[4166038]: segfault at 80 ip
7effd99c3964 sp 7ffee513d5f8 error 4 in
libnftnl.so.11.2.0[7effd99be000+19000]
[Fri Nov 12 11:43:46
Package: unscd
Version: 0.54-1
The unscd.prerm script should cleanup the caches
nscd -i {group,hosts,netgroup,passwd,services}
rm -f /var/cache/nscd/{group,hosts,netgroup,passwd,services}
to invalidate and remove all caches, before the service is stopped
and removed.
The nscd
Package: nscd
Version: 2.31-13+deb11u2
The nscd.prerm script should cleanup the caches
nscd -i {group,hosts,netgroup,passwd,services}
rm -f /var/cache/nscd/{group,hosts,netgroup,passwd,services}
to invalidate and remove all caches, before the service is stopped
and removed.
metoo
This bug triggers a bazillion EMails for cron jobs that are usually
very quiet, except for a catastrophe.
Please fix.
Thank you very much
Harri
Package: apt
Version: 2.3.9
Severity: wishlist
Unfortunately there is a pretty high risk for package name conflicts,
if multiple package repositories are involved. There is no reliable
way to make sure that some 3rd-party repo doesn't provide its own
"improved" openssl packages, for example.
PS: Using sysvinit both services are fine.
PS: I should have mentioned that the nvidia kernel package of 470.57.02-2
doesn't build for kernel 5.14.
Package: aptitude
Version: 0.8.11-7
If I add minio Version 20210805220119.0.0 to a private package
repository, then aptitude succeeds on the first run. Press 'u'
to update the Packages file. On the second run aptitude dies
with SIGSEGV. After removing the cached Packages file listing minio &
Package: nvidia-kernel-dkms
Version: 470.57.02-2
Since upstream's kernel 5.13 has reached EOL and since nvidia-graphics-drivers
470.63.01 seems to build fine with 5.14 (http://rglinuxtech.com/?cat=18) I
wonder
if it would be possible to provide the new version for Sid?
Regards
Harri
Sorry, I cannot reproduce this anymore. The auto.misc.ucf-old has been deleted.
Regards
Harri
Package: nfs-kernel-server
Version: 1:1.3.4-6
After boot the nfs-server is not ready yet:
# systemctl status nfs-server
* nfs-server.service - NFS server and services
Loaded: loaded (/lib/systemd/system/nfs-server.service; enabled; vendor
preset: enabled)
Active: failed (Result:
PPS: Using ftp2.de.debian.org instead of deb.debian.org there
is no such problem.
% host deb.debian.org
deb.debian.org is an alias for debian.map.fastlydns.net.
debian.map.fastlydns.net has address 151.101.14.132
debian.map.fastlydns.net has IPv6 address 2a04:4e42:3::644
% host
I've got a similar problem, using repositories on deb.debian,org.
'apt update' fails to obtain the InRelease files from acng:
:
Ign:18 http://HTTPS///pkg.jenkins.io/debian binary/ InRelease
Hit:19 http://HTTPS///pkg.jenkins.io/debian binary/ Release
Err:21 http://deb.debian.org/debian bullseye
PS: acng is version 3.6.4-1
PS: AFAICT its sufficient to include a line
/var/log/charon.log rw
and to run
apparmor_parser -r /etc/apparmor.d/usr.lib.ipsec.charon
Regards
Harri
Package: strongswan-charon
Version: 5.9.1-1
Severity: wishlist
Would you mind to add some sample code to usr.lib.ipsec.charon to
support logging, e.g. for writing to /var/log/charon.log ?
Currently its much easier to kick out apparmor than to figure out
how to extend usr.lib.ipsec.charon to
Package: meld
Version: 3.20.3-1
All it says is
# meld /etc/auto.misc.ucf-old /etc/auto.misc
Trace/breakpoint trap
Package: lego
Version: 3.2.0-3.1+b5
Upstream provides a new version lego 4.4.0, see
https://github.com/go-acme/lego
Would it be possible to include this new version into Debian?
Thanx in advance
Harri
Package: lxc
Version: 1:4.0.6-2
Severity: wishlist
Upstream provides a new version 4.0.10 and 4.0.9LTS, see
https://discuss.linuxcontainers.org/t/lxc-4-0-10-has-been-released/11618
https://discuss.linuxcontainers.org/t/lxc-4-0-9-lts-has-been-released/10999
No build problems
Using 470.42.01-1 and upstream's kernel 5.13.2 the problem seems to be gone.
Hi Julien,
I'm fine with keeping the code as it is. It was just weird to
find dpkg's file locking come up for something that should not
involve dpkg at all.
Of course I found debget, so I am happy. Ticket can be closed.
Regards
Harri
Package: apt
Version: 2.2.4
Trying to download a package using an unprivileged account fails with
% apt -o dir::cache::archives="/tmp/" -d install mg
E: Could not open lock file /var/lib/dpkg/lock-frontend - open (13: Permission
denied)
E: Unable to acquire the dpkg frontend lock
Second that, except that the current version is 2.8.0 now.
Regards
Harri
Upstream seems to provide Debian support:
https://github.com/evilsocket/opensnitch/tree/master/debian
Package: azure-cli
Version: 2.18.0-2
"az aks update" died with
module.cluster.null_resource.cluster (local-exec): Executing: ["/bin/sh" "-c" "az
aks update --resource-group $RESOURCE_GROUP --name $CLUSTER_NAME --attach-acr $REGISTRY_ID"]
module.cluster.null_resource.cluster (local-exec):
metoo
# ras-mc-ctl --summary
No Memory errors.
No PCIe AER errors.
No Extlog errors.
DBD::SQLite::db prepare failed: no such table: devlink_event at
/usr/sbin/ras-mc-ctl line 1183.
Can't call method "execute" on an undefined value at /usr/sbin/ras-mc-ctl line
1184.
# ras-mc-ctl --errors
Hi Eduard,
I have to admit I haven't thought about broken ssl configurations
and diverging folder configurations for https. Its fine to reject
this feature request.
Thank you very much, anyway.
Harri
I could reproduce this using linux-image-5.10.0-7-amd64.
This thread seems to be related:
https://forums.developer.nvidia.com/t/465-24-02-page-fault/175782/31
Could you please keep 460.73.01-1 in Testing?
Regards
Harri
Package: logrotate
Version: 3.18.1-1
Severity: wishlist
Would you mind to enable logrotate logging for Bookworm by default?
logrotate appears to be excellent in logging. Its a shame that this
is not used. And since you have to wait 24h for the next run (unless
you enable hourly rotation) it is
Package: apt-cacher-ng
Version: 3.6.3-1
Severity: wishlist
Sorry to say, but configuring https for apt-cacher-ng is APITA. Would it be
possible for ACNG to silently try https first, if the client asked for http?
That would be similar to an explicit http://HTTPS///get.docker.com/ubuntu,
except
Package: nvidia-kernel-dkms
Version: 460.80-1
I've got a reproducible freeze for kernel 5.12.6 (built from git) and
the new 460.80-1 nvidia packages. Instead of a graphics screen I just
get a cursor in the top left corner. The console is unresponsive. I
cannot switch to another console. The Xorg
We go in circles here.
651236c2319eb0ca13fd1d376eaf239a6dcd5c49 has been introduced as a
workaround, because logrotate/rsyslog was losing SIGHUPs in the
previous configuration. This special configuration introduced
with 651236c2319eb0ca13fd1d376eaf239a6dcd5c49 is something I cannot
use on some
On 5/19/21 12:10 PM, Michael Biebl wrote:
Does that mean, you manually modified /etc/logrotate.d/rsyslog to have
multiple sections / postrotate calls?
Of course I have to modify the default settings on some of my hosts.
I did not modify /etc/logrotate.d/rsyslog to have multiple sections.
It
Hi Michael,
On 5/18/21 6:42 PM, Michael Biebl wrote:
On Thu, 25 Feb 2021 13:30:40 +0100 Harald Dunkel
My favorite workaround is to kick out the delaycompress, see
https://bugs.debian.org/720096#130.
I still fail to see a convincing explanation, why that would fix this issue.
Its
Hi David,
the fix is a pretty obvious one-liner. The risk in applying it is very
low. But it makes the difference between "works for me" and "better
don't use opendkim, because it makes things worse".
Consider this: You send an EMail to somebody outside your own (masqueraded)
domain, you get a
Hi David,
Would you mind to apply the patch to opendkim in Sid to increase
the chance it can make it into Bullseye? Of course I had verified
it on my MTA using a private backport of 2.11.0~beta2-4 to Buster.
Thanx very much
Harri
- The unit opendkim.service has entered the 'failed' state with result
'exit-code'.
:
See https://sourceforge.net/p/opendkim/bugs/257/ for a fix (attached).
This is important. Without this feature opendkim is not usable
together with sendmail, if masquerading is enabled.
--
Dipl.-Ing. Har
On Mon, 22 Feb 2021 20:00:33 +0100 Michael Biebl wrote:
Am 22.02.2021 um 18:57 schrieb Harald Dunkel:
> Sorry to say, but this is not a fix. A fix would avoid the race
> condition, no matter whats written in the config files.
>
> Your "fix" is just a workaround. If I add
It is still broken. I have installed the new version on our mail
server yesterday. This morning I found this:
root@il01:~# ls -alrt /var/log/daemon.log*
-rw-r- 1 root adm 5818753 Jan 31 00:00 /var/log/daemon.log.5.gz
-rw-r- 1 root adm 5841733 Feb 7 00:00 /var/log/daemon.log.4.gz
Sorry to say, but this is not a fix. A fix would avoid the race
condition, no matter whats written in the config files.
Your "fix" is just a workaround. If I add
/var/log/logfile1
{
rotate 7
daily
missingok
notifempty
delaycompress
compress
Actually I cannot say. I haven't seen this problem in a while. And
/etc/init.d/nfs-common start does
PIPEFS_MOUNTPOINT=/run/rpc_pipefs
:
mkdir -p "$PIPEFS_MOUNTPOINT"
:
modprobe sunrpc nfs nfsd
mount -t rpc_pipefs rpc_pipefs $PIPEFS_MOUNTPOINT
Package: openssh-server
Version: 1:8.4p1-3
Severity: wishlist
sshd works very well without /etc/ssh/sshd_config, using just
the defaults, so I wonder if the openssh-server package should
provide /etc/ssh/sshd_config at all?
Providing and maintaining a (sparse) /etc/ssh/sshd_config could
be the
PS: The commit on LXC's stable-4.0 branch is
563ec46266b8967f0ee60e0032bbe66b3b37207c
Regards
Harri
Package: lxc
Version: 1:4.0.6-1
563ec46266b8967f0ee60e0032bbe66b3b37207c introduced a bug into
lxc 4.0.6: /proc/sys/net is mounted read-only, even though the
container is privileged and did not drop CAP_NET_ADMIN. Result:
# echo 0 >/proc/sys/net/ipv4/ip_forward
bash:
Package: rasdaemon
Version: 0.6.6-2
rasdaemon doesn't start. If I try to run it manually, then there is
an error message
# rasdaemon --foreground
rasdaemon: Can't find debugfs
rasdaemon: Can't locate a mounted debugfs
# type debugfs
debugfs is
Package: zstd
Version: 1.3.8+dfsg-3
Severity: critical
Compressing a large file with restricted access permissions a new,
world readable file is created, revealing the contents of the
uncompressed file. Sample:
# whoami
root
# zstd -q -13 -T8 sample.dmp &> zstd.log &
:
:
# ls -al
total
Package: minidlna
Version: 1.3.0+dfsg-1
Tags: patch
Apparently minidlna is started before autofs and others. This
breaks the initial scan for new media files. On the next reboot
you just get an empty database.
Attached is a patch for minidlna.init and minidlna.service.
Regards
Harri
---
Log file is attached.
As you can see, Docker is involved. AFAICT that was Docker version
5:20.10.2~3-0~debian-buster, provided by upstream as a Debian package.
The host was on Sid.
In the meantime this host has been reconfigured to run Kubernetes without
Docker (WIP).
Regards
Harri
AFAICS the host ran into OOM.
Regards
Harri
Package: grub-pc
Version: 2.04-12
The grub-pc.postinst script lists all my disks to install grub. It
also provides an option to install grub on /dev/md0 (i.e. my RAID1),
even though there is no valid boot sector. Its a physical volume for
LVM2. It is easy to select /dev/md0 in the debconf menu
As a workaround, have you tried to boot with
systemd.unified_cgroup_hierarchy=0
?
Package: lxc
Version: 1:4.0.5-2
Upstream provided version 4.0.6 of lxc, see
https://github.com/lxc/lxc/releases/tag/lxc-4.0.6
Would be nice if the new version could make it into Bullseye.
Regards
Harri
Package: systemd
Version: 247.2-4
I do not know how my host got into this state
root@srvl065:~# systemctl status
Failed to read server status: Transport endpoint is not connected
but AFAICT this is not supposed to happen. Google doesn't provide a
recovery procedure, i.e. this
101 - 200 of 1834 matches
Mail list logo
