Problem with installation on Debian 13, computer hangs after several hours

[JohannesE]

I am running VDR on a debian server for many years.
Last year i bought an Sundtek SkyTV Ultimate Dual UHD and it worked without problems.

After an upgrade to debian 13, the server is working unstable, it hangs after several hours.
Because the server was upgraded over several debian versions, i decided to do a fresh install on a new SSD on the same computer, but i get the same behavior.

To examin the problem, I checked the logfiles, in dmesg i found messages like these:

[Jan24 12:42] audit: type=1400 audit(1769254930.432:150): apparmor="DENIED" operation="open" class="file" profile="kea-dhcp4" name="/opt/lib/libmediaclient.so" pid=1013100 co>
[Jan24 13:17] audit: type=1400 audit(1769257021.913:151): apparmor="ALLOWED" operation="open" class="file" profile="unix-chkpwd" name="/opt/lib/libmediaclient.so" pid=1020843>
[  +0,000014] audit: type=1400 audit(1769257021.913:152): apparmor="ALLOWED" operation="file_mmap" class="file" profile="unix-chkpwd" name="/opt/lib/libmediaclient.so" pid=10>
[Jan24 13:30] audit: type=1400 audit(1769257801.376:153): apparmor="ALLOWED" operation="open" class="file" profile="unix-chkpwd" name="/opt/lib/libmediaclient.so" pid=1023718>
[  +0,000006] audit: type=1400 audit(1769257801.376:154): apparmor="ALLOWED" operation="file_mmap" class="file" profile="unix-chkpwd" name="/opt/lib/libmediaclient.so" pid=10>

After some investigation i found the file the file /etc/ld.so.preload, which loads /opt/lib/libmediaclient.so.
For testing, i removed this file and without it, i get no longer server hang-ups and the messages in demsg also have disappeared.
But without this preload, VDR is no longer working.

Is there any other way to load this shared library? It should only be loaded for programs that require it.

The VDR version, which is included in the stable debian version is 2.6.9-3. Could this be a problem?

The VDR server has no graphical desktop.
When the server hangs, it responds to a network ping, but it is not possible to login to the server with ssh or on a local console.
But i can login to local consoles in advance and run tools like dmesg, htop or similar. Then, when the server hangs, i still can switch between these consoles (Alt-F1, Alt-F2...) and read the output of these tools.

What can i do to get the server running stable?

[Sundtek]

So basically just ssh is blocked? Possibly there are some apparmor items which block it, can you check for other items in the logfile which show something else than ALLOWED?
Possibly the new distribution came with some new rules on board which interfere. Maybe you can join the discord chat so we can chat about your problem (that might be fastest)?

You have a few options at that point (to get VDR work with it).

The discord chat is here:
https://discord.gg/AR5QJhyNnw

I have installed Trixie on our side, but cannot reproduce any problem - so I will need your help to see what's going on for you.

[JohannesE]

Thanks for your reply. Yesterday i was not at home, so i could not answer.

i did a little more testing:
I restored the ld.so.preload file and stopped VDR with "systemctl stop vdr"; the server is running now for nearly 24 h and didn't hang.

I still get the apparmor warnings, most logfile entries are with "ALLOWED", the only message that contains "DENIED" is this one (it repeates every 60 minutes):
audit: type=1400 audit(1769459579.441:264): apparmor="DENIED" operation="open" class="file" profile="kea-lfc" name="/opt/lib/libmediaclient.so" pid=306381 comm="kea-lfc" requested_mask="r" denied_mask="r" fsuid=115 ouid=0

So maybe the server hangup is more a problem with VDR than the USB driver.
Did you also run VDR on the debian trixie? How long did you let it run? Here the server sometimes runs for several hours or half a day before it gets unstable, sometimes it happens very quickly.

I think i will do some more testing without any VDR plugins, just a minimal installation and maybe i will test a more recent VDR version.
What logfiles would be interesting to examine the problem further?

[Sundtek]

with "hang" you mean the network connectivity problem?

The DENIED message should be fine because kea-lfc doesn't need libmediaclient.so.

You can always isolate the LD_PRELOAD to VDR only by modifying the vdr systemd startscript.

It's difficult to have the same setup as you do, so I'd prefer to isolate the problem on your side to fix your problem.

[JohannesE]

> with "hang" you mean the network connectivity problem?

It's not just a network connectivity problem.
It is also not possible to login on a local console.

I can log in locally or by SSH over network, before the server hangs.
Then i wait until it hangs; when i try to execute a command, e.g. "ls", nothing happens at first, i get no response. After pressing <Ctrl>-C, i get the output of the ls-command.
Starting of other programs is also not possible.

Also, the 'poweroff'-command does not work. The server starts to shut down, but then at some point it hangs and can only be restarted be a a power cycle.

Yesterday in the evening, i uninstalled all VDR plugins except vdr-plugin-vnsiserver. Then i connected with KODI from a different computer and did some testing.

Watching television worked. I started a recording and tried to watch a recorded movie while another one was recorded. In this situation, the problem happened again and i had to restart the server.
It seems, the error is triggered by some actions from VDR, but it also happens after some time when i only startup the server with no KODI client connected.

[Sundtek]

This sounds like a driver / kernel problem.

- Check the system logs.
- Also check the memory utilisation during runtime.

I highly doubt that this has anything to do with our driver.

The good thing seems to be that it is reproducible, the closer you can narrow it down workarounds or other fixes could be applied.

From your description it sounds like a problem with the SSD or controller for it.

[JohannesE]

I have an update:

I deleted the /etc/ld.so.preload-file and added a file override.conf to /etc/systemd/system/vdr.service.d/ and /etc/systemd/system/sundtek.service.d

The override.conf file contains these two lines:
[Service]
Environment=LD_PRELOAD=/opt/lib/libmediaclient.so

I removed all VDR plugings and other packages related to VDR, except vdr and vdr-plugin-vnsiserver.

With this configuration, the server runs stable and VDR works; i can connect with KODI from a remote PC.
It was running for more than 36 hours without any problem.

Then i restored the file /etc/ld.so.preload which contains only one line "/opt/lib/libmediaclient.so" and executed "systemctl daemon-reload" (no reboot); after a few hours, the same problem happened again.

For debugging, i logged into all 6 local tty-consoles with user "root" on the server directly after booting.
In one console, i ran "htop", in another console i ran "dmesg -w" and some other logs in other consoles.

After a few hours, i noticed the server had the problem. During this time, VDR was in idle mode, no KODI was runnig.
On the local console, the htop output was freezed, but it showed no high cpu or memory usage. In the dmesg output, i had a lot of apparmor warnings with name="/opt/lib/libmediaclient.so", similar like in my first post.
I can switch between these consoles with <Alt>-F1/2/3...
When i enter a command in a console, e.g., "ls", it doesn't respond. But then, after pressing <CTRL>-C, i get the output of the ls command.

Then i deleted the file /etc/ld.so.preload and executed "systemctl daemon-reload", always with <CTRL>-C afterwards. After doing this, the server seemed to operate normal.
I tried to restart VDR with "systemctl restart vdr", but this command did not finish, i had to press <CTRL>-C.
I tried it multiple times and also with stop and start separately.

Then i did a reboot, it took a long time to shut down (several minutes).
After rebooting, i checked the VDR log (journalctl -u vdr).

There you can see some error messages from the failed vdr start before reboot:

Jan 30 13:26:00 srv1 systemd[1]: Starting vdr.service - Video Disk Recorder...
Jan 30 13:26:00 srv1 vdr[35811]: [35811] VDR version 2.6.9 started
Jan 30 13:26:00 srv1 vdr[35811]: [35811] switched to user 'vdr'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] codeset is 'UTF-8' - known
Jan 30 13:26:00 srv1 vdr[35811]: [35811] found 28 locales in /usr/share/locale
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'alb,sqi'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'bos'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'bul'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'chi,zho'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'eus,baq'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'iri,gle'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'jpn'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'lav'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'ltz'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'mlt'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'por'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no locale for language code 'smi'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading plugin: /usr/lib/vdr/plugins/libvdr-vnsiserver.so.2.6.9
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/setup.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/sources.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/diseqc.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/scr.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/channels.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/timers.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/commands.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/reccmds.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/svdrphosts.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35811] loading /var/lib/vdr/keymacros.conf
Jan 30 13:26:00 srv1 vdr[35811]: [35813] video directory scanner thread started (pid=35811, tid=35813, prio=low)
Jan 30 13:26:00 srv1 vdr[35811]: [35814] epg data reader thread started (pid=35811, tid=35814, prio=high)
Jan 30 13:26:00 srv1 vdr[35811]: [35811] registered source parameters for 'A - ATSC'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] registered source parameters for 'C - DVB-C'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] registered source parameters for 'S - DVB-S'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] registered source parameters for 'T - DVB-T'
Jan 30 13:26:00 srv1 vdr[35811]: [35811] detected /dev/dvb/adapter1/frontend0
Jan 30 13:26:00 srv1 vdr[35811]: [35811] detected /dev/dvb/adapter0/frontend0
Jan 30 13:26:00 srv1 vdr[35811]: [35814] reading EPG data from /var/cache/vdr/epg.data
Jan 30 13:26:00 srv1 vdr[35811]: [35811] ERROR (dvbdevice.c,1945): /dev/dvb/adapter0/frontend0: Kein passendes Gerät bzw. keine passende Adresse gefunden
Jan 30 13:26:00 srv1 vdr[35811]: [35811] ERROR (dvbdevice.c,1945): /dev/dvb/adapter1/frontend0: Kein passendes Gerät bzw. keine passende Adresse gefunden
Jan 30 13:26:00 srv1 vdr[35811]: vdr: no primary device found - using first device!
Jan 30 13:26:00 srv1 vdr[35811]: [35811] no DVB device found
Jan 30 13:26:00 srv1 vdr[35811]: [35811] initializing plugin: vnsiserver (1.8.3): VDR-Network-Streaming-Interface (VNSI) Server
Jan 30 13:26:00 srv1 vdr[35811]: [35811] ERROR: invalid primary device number: 1
Jan 30 13:26:00 srv1 vdr[35811]: [35811] ERROR: no primary device found - using first device!
Jan 30 13:26:00 srv1 vdr[35811]: [35811] ERROR: invalid primary device number: 1
Jan 30 13:26:00 srv1 vdr[35811]: [35813] video directory scanner thread ended (pid=35811, tid=35813)
Jan 30 13:26:00 srv1 vdr[35811]: [35814] epg data reader thread ended (pid=35811, tid=35814)
Jan 30 13:26:01 srv1 vdr[35811]: [35811] deleting plugin: vnsiserver
Jan 30 13:26:01 srv1 vdr[35811]: [35811] max. latency time 0 seconds
Jan 30 13:26:01 srv1 vdr[35811]: [35811] exiting, exit code 2
Jan 30 13:26:01 srv1 systemd[1]: vdr.service: Main process exited, code=exited, status=2/INVALIDARGUMENT
Jan 30 13:26:01 srv1 systemd[1]: vdr.service: Failed with result 'exit-code'.
Jan 30 13:26:01 srv1 systemd[1]: Failed to start vdr.service - Video Disk Recorder.
Jan 30 13:26:01 srv1 systemd[1]: vdr.service: Consumed 864ms CPU time, 121.7M memory peak.


So for me i have a solution, but maybe it would be interesting to find out why the servers hangs when the ld.so.preload file is installed.
Do you have any ideas, which logfiles i could check?

[Sundtek]

Check the system log

dmesg
/var/log/kern.log
/var/log/syslog

run strace on processes which appear to hang. To me it looks like a SSD problem in some way, we have an old Mac which behaves like that with a defect SSD. (of course I migth be wrong but that's just my assumption right now).

Maybe check the ssd with smarttools

eg.
sudo smartctl -a /dev/sda
VDR is probably started before the driver is fully initialized.

To let the driver restart VDR after all devices are initialized:
/etc/sundtek.conf
device_attach=systemctl restart vdr
bulknotification=on

if the problem happens you can join the discord chat and might discuss which logs to obtain via chat, that might be faster than via slow forum ping pong.
Here again the discord/chat invitation link:
https://discord.gg/AR5QJhyNnw