So for the past couple of days I've been getting pinged by Fix Common Problems about OOM errors and have come back to find my AMP server VM shut off. It came to a head the other day when while I was watching the Dashboard, I saw my in use memory shoot from ~42GB (of 64GB) to completely maxed out for a moment before unRAID locked up entirely, requiring a hard reset.

Now, that AMP VM is set to start at 8GB and peak at 32GB, so it becomes the prime target for the OOM Killer.

At first I thought maybe transcoding to /dev/shm was the problem, so I've temporarily set Plex to transcode to my cache.

Today it OOM'd twice, the first time killing the VM again, but the second time it killed the Plex Transcoder, and if I'm reading this right:

Out of memory: Killed process 4127540 (Plex Transcoder) total-vm:55839484kB, anon-rss:55779032kB, file-rss:364kB, shmem-rss:0kB, UID:99 pgtables:109296kB oom_score_adj:0

It was using ~53 GB of physical RAM when it was killed. If I'm misreading that do tell me, I'm still fairly new to unRAID and not nearly as well versed in Linux as I am in Windows.

EDIT: For anyone who finds this thread later on, the issue was a corrupted Blobs DB. The DBRepair script was able to set things straight and the transcoder process is no longer causing problems.

Diagnostics attached.

raidcore-diagnostics-20251124-2034.zip

Edited November 26, 2025Nov 26 by CMDMC12

7 hours ago, CMDMC12 said:

So for the past couple of days I've been getting pinged by Fix Common Problems about OOM errors and have come back to find my AMP server VM shut off. It came to a head the other day when while I was watching the Dashboard, I saw my in use memory shoot from ~42GB (of 64GB) to completely maxed out for a moment before unRAID locked up entirely, requiring a hard reset.

Now, that AMP VM is set to start at 8GB and peak at 32GB, so it becomes the prime target for the OOM Killer.

At first I thought maybe transcoding to /dev/shm was the problem, so I've temporarily set Plex to transcode to my cache.

Today it OOM'd twice, the first time killing the VM again, but the second time it killed the Plex Transcoder, and if I'm reading this right:

Out of memory: Killed process 4127540 (Plex Transcoder) total-vm:55839484kB, anon-rss:55779032kB, file-rss:364kB, shmem-rss:0kB, UID:99 pgtables:109296kB oom_score_adj:0

It was using ~53 GB of physical RAM when it was killed. If I'm misreading that do tell me, I'm still fairly new to unRAID and not nearly as well versed in Linux as I am in Windows.

Diagnostics attached.

raidcore-diagnostics-20251124-2034.zip

You've got it right - make sure it's actually transcoding to disk, and if all else fails you can limit the memory allowed to the container.

It could be choking on a particular file, too - judging by how it spikes rapidly (I've had that happen with Plex as well). So you can investigate that angle by narrowing down any recent additions to the library

4 hours ago, Michael_P said:

You've got it right - make sure it's actually transcoding to disk, and if all else fails you can limit the memory allowed to the container.

It could be choking on a particular file, too - judging by how it spikes rapidly (I've had that happen with Plex as well). So you can investigate that angle by narrowing down any recent additions to the library

Yep, I force transcoded a movie to verify it was actually on disk now, and good call on adding the memory limit. Set the container to 8GB and overnight the transcoder got killed 4 separate times in a span of 10 minutes, but it no longer takes down my VM so I consider that a win for now.

Now I guess I need to dig through Plex's logs to narrow things down.