rmccullough
Patron
- Joined
- May 17, 2018
- Messages
- 269
I am running Lidarr (like Sonarr/Radarr but for music) in an iocage jail on FreeNAS 11.2 U1. Over the last 2 months I have observed the process hang 2-3 times where the Web UI is unresponsive and the mono/Lidarr.exe process is using 100% of a CPU. Restarting the service is stuck waiting for the PID to end. Performing a "kill -9" does not kill the process. Attempting to reboot the FreeNAS server also hangs with the system unable to end the process. I am forced to hard reset the system.
I opened a case against Lidarr on their Github page: https://github.com/lidarr/Lidarr/issues/602
They asked for a "truss" log against the mono PID when the issue occurs. Since I cannot reproduce this, I don't know how to obtain that. For fun I ran the "truss" command and provided some output here: https://pastebin.com/xnnbrt4S
This is their response:
Is this a known problem? Should I go log a ticket on redmine?
I opened a case against Lidarr on their Github page: https://github.com/lidarr/Lidarr/issues/602
They asked for a "truss" log against the mono PID when the issue occurs. Since I cannot reproduce this, I don't know how to obtain that. For fun I ran the "truss" command and provided some output here: https://pastebin.com/xnnbrt4S
This is their response:
Yes, that is extremely helpful. The pointer here is a flood of UMTX_OP_WAIT_UINT_PRIVATE going on, indicating the process is spinning uselessly trying to get a lock in kernel space. Most likely related to a file operation or high rate of file operations. This does NOT reproduce on FreeBSD 11.2 or 12.0 with ZFS.
There are complaints about numerous other applications that access files frequently running into the same problems, making it clearly Yet Another FreeNAS Specific Issue
Is this a known problem? Should I go log a ticket on redmine?