Hi all,
I am lacking experience but I don't think this is only a Jellyfin related issue as my whole homelab server crashes.
When I start Jellyfin video on my 4K HDR TV the whole server will crash after only a few seconds to max 3 minutes. The stream will suddenly stop working and the Proxmox Web UI will become unresponsive. The whole server does not react anymore. When I stream the same video from my computers Browser (WQHD resolution), everything works fine. At a crash, I have to fully shutdown the system by holding down the on/off button and restart.
I have one of those simple homelab firewall mini computers with an Intel N100 (bought it barebone), a Kingston NV3 nvme ssd for everything (only storage in the system) and 16GB of RAM running proxmox VE 8.3.3. Jellyfin runs as a privileged LXC with the GPU being passed through (Intel N100 GPU). Jellyfin is set up according to their docs. In Jellyfin I set up transcoding to what the GPU should be able to handle and everything seemed to work smoothly. Transcoding works, the cpu load stays low (below 10% usually), RAM usage stays low (couple hundred MB), temps stay low (~45°C), everything seems fine. The only thing that happens what I noticed is an increasing IO wait when transcoding to 4K is active, the IO wait rises to over 30%. I noticed that my ssd might not be the best as it has bad caches or so, but that does not explain the crash, right? Maybe my RAM is bad, but wouldn't I get errors then in the journal log?
The journal logs do not show an error:
Also the Jellyfin logs show nothing. When the crash occures, it only shows
I attached files for the nodes and the containers summary view at crash. The temps in my system never exceeded 55°C, neither for the CPU, nor the ssd.
How can I figure out what really happens?
Thanks for all your help.
I am lacking experience but I don't think this is only a Jellyfin related issue as my whole homelab server crashes.
When I start Jellyfin video on my 4K HDR TV the whole server will crash after only a few seconds to max 3 minutes. The stream will suddenly stop working and the Proxmox Web UI will become unresponsive. The whole server does not react anymore. When I stream the same video from my computers Browser (WQHD resolution), everything works fine. At a crash, I have to fully shutdown the system by holding down the on/off button and restart.
I have one of those simple homelab firewall mini computers with an Intel N100 (bought it barebone), a Kingston NV3 nvme ssd for everything (only storage in the system) and 16GB of RAM running proxmox VE 8.3.3. Jellyfin runs as a privileged LXC with the GPU being passed through (Intel N100 GPU). Jellyfin is set up according to their docs. In Jellyfin I set up transcoding to what the GPU should be able to handle and everything seemed to work smoothly. Transcoding works, the cpu load stays low (below 10% usually), RAM usage stays low (couple hundred MB), temps stay low (~45°C), everything seems fine. The only thing that happens what I noticed is an increasing IO wait when transcoding to 4K is active, the IO wait rises to over 30%. I noticed that my ssd might not be the best as it has bad caches or so, but that does not explain the crash, right? Maybe my RAM is bad, but wouldn't I get errors then in the journal log?
The journal logs do not show an error:
Code:
root@proxmox:~# journalctl --since "2025-02-03 18:38:00" --until "2025-02-03 18:50:00"
Feb 03 18:40:31 proxmox pvedaemon[971]: <root@pam> successful auth for user 'root@pam'
Feb 03 18:40:31 proxmox pvedaemon[414237]: starting termproxy UPID:proxmox:0006521D:006F10A0:67A0FF8F:vncshell::root@pam:
Feb 03 18:40:31 proxmox pvedaemon[973]: <root@pam> starting task UPID:proxmox:0006521D:006F10A0:67A0FF8F:vncshell::root@pam:
Feb 03 18:40:31 proxmox pvedaemon[972]: <root@pam> successful auth for user 'root@pam'
Feb 03 18:40:31 proxmox login[414240]: pam_unix(login:session): session opened for user root(uid=0) by (uid=0)
Feb 03 18:40:31 proxmox systemd-logind[630]: New session 184 of user root.
Feb 03 18:40:31 proxmox systemd[1]: Started session-184.scope - Session 184 of User root.
Feb 03 18:40:31 proxmox login[414246]: ROOT LOGIN on '/dev/pts/1'
Feb 03 18:40:33 proxmox systemd[1]: session-184.scope: Deactivated successfully.
Feb 03 18:40:33 proxmox systemd-logind[630]: Session 184 logged out. Waiting for processes to exit.
Feb 03 18:40:33 proxmox systemd-logind[630]: Removed session 184.
Feb 03 18:40:33 proxmox pvedaemon[973]: <root@pam> end task UPID:proxmox:0006521D:006F10A0:67A0FF8F:vncshell::root@pam: OK
-- Boot ea8278a6924c4137a8b4cd2dee1028c4 --
Feb 03 18:46:48 proxmox kernel: Linux version 6.8.12-7-pve (build@proxmox) (gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40) #1 SMP PREEMPT_DYNAMIC PMX 6.8.12>writing����������������I attached files for the nodes and the containers summary view at crash. The temps in my system never exceeded 55°C, neither for the CPU, nor the ssd.
How can I figure out what really happens?
Thanks for all your help.
