Docker Swarm + Unifi Switch = Massive Packet Loss?

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

nitrobass24

Moderator
Dec 26, 2010
1,087
131
63
TX
Ran all weekend best I could tell without an issue then I try to stream a TV show and have issues. Shutdown the sonarr container and the issues persisted.

Moving down the stack. I am basically left with an ESXi or FreeNAS issue.
 

nitrobass24

Moderator
Dec 26, 2010
1,087
131
63
TX
Well it turns out its my FreeNAS storage system. NFSd just stops/starts for no reason, but fails.

Here is the tail of the /var/log/messages

Code:
[root@freenas] ~# cat /var/log/messages | grep nfs
May 30 00:05:24 freenas notifier: Stopping nfsd.
May 30 00:05:24 freenas notifier: Stopping nfsuserd.
May 30 00:05:25 freenas notifier: Starting nfsuserd.
May 30 00:05:25 freenas notifier: Starting nfsd.
May 30 00:05:25 freenas nfsd[60563]: Can't read stable storage file: Operation not permitted
May 30 00:08:22 freenas notifier: nfsd not running? (check /var/run/nfsd.pid).
May 30 00:08:22 freenas notifier: Stopping nfsuserd.
May 30 00:08:22 freenas notifier: Starting nfsuserd.
May 30 00:08:22 freenas notifier: Starting nfsd.
May 30 00:08:23 freenas nfsd: can't register svc name
May 30 00:36:34 freenas nfsd: can't register svc name
[root@freenas] ~#
WTF
 

nitrobass24

Moderator
Dec 26, 2010
1,087
131
63
TX
Updated FreeNAS to 9.10 U4 and still have this issue. Definitely, something with FreeNAS, because when it happens I lose all NFS connectivity everywhere simultaneously.
 

nitrobass24

Moderator
Dec 26, 2010
1,087
131
63
TX
No not really any closer to figuring this out, unfortunately. I have completely rebuilt my VMware Cluster and while it does not happen on any regular basis, I have seen it occur once since.

Considering I lose NFS connectivity on both ESXi host simultaneously I am basically down to some sort of Switch Issue or a FreeNAS issue. I am leaning FreeNAS because I didn't have this issue on my Synology, but I, unfortunately, cant rule out the Switch.

I might buy two Small SSDs to run the Docker workload and put them back into my Syno and see if I can re-produce the issue.
 

nitrobass24

Moderator
Dec 26, 2010
1,087
131
63
TX
UBNT has agreed to replace my switch after looking at my logs and .supp files for a few weeks. "They would like the switch back for physical inspection".