Any guess to where my starting point for troubleshooting should be.
Basic setup is:
Freenas baremetal where all media is stored
Unifi switch as networking component
Esxi 6.7 host where VM’s are hosted
Fedora VM which hosts docker containers amongst which is Plex
I can pretty much “exclude” every component
Other containers on the same VM are running perfectly fine ans also using the Freenas volume.
As such it would stand to reason the Freenas host is OK, the network is OK, the ESXi host is OK, the Fedora VM is OK.
And with nothing in the Plex logs, and no-one else experiencing this issue, the Plex container is probably also OK… so it must be a combination of factors at play, but I don’t even know where to start…
Several posts about crashing before switching VMXNET3 to E1000e and turning of LRO. Thought I saw another thread about creating the vm with 6.0 but can’t find that right now.
So maybe I found something…
I switched from NFS to CIFS as a test, and (fingers crossed) it seems to be more stable…
The behaviour is all pretty logical if you consider how NFS works…
From Plex not actually crashing but just becoming entirely unresponsive… to the high iowait on the CPU’s, to Plex returning after restarting the storage server…
It’s because NFS just waits… instead of timing out like cifs/smb
Do I’ll update in a few days if this keeps on working…
I don’t have enough know-how to fix NFS, but if this workaround works, it’s a good second choice
3 days later and my issue is non-existant anymore…
For some reason NFS was causing the issue
As a workaround de CIFS/SMB share works perfectly, no more hangsups.
Strangely, only Plex was hanging, even though more containers use the same NFS connection, and evene multiple devices… none of them have the issue…
