Odd NFS/Network issues

TrueNAS General
1

I have 6 larger truenas boxes setup nearly identical. 4 working flawlessly. the 5th worked flawlessly under 24.10.1. When updated to the current 25.04.2.4 I get some really weird issues:
kernel: loop0: detected capacity change from 0 to 3273944
systemd-journald[1333]: /var/log/journal/cb81a0efb3b4492b9b7bb466e4a783f5/system.journal: Journal header limits reached or header out-of-date, rotating.
kernel: rpc-srv/tcp: nfsd: sent 381040 when sending 1048704 bytes - shutting down socket

The other 4 servers work perfectly fine on the current version.

All servers run the same motherboard. Dual E5-2697 V2, 4 have 1tb of ram, 2 have 512gb of ram. they all run 12gb sas conntroller, 24hd drives, they all run intel p3700 mirrored for SLOG, they all have 40gb Chelsios Cards, the only difference is how the L2Arc is done. the 4 have 12 800gb sas drives to a dedicated sas controller, and the 2 have 2 6.4gb pci Intel NVMe controllers, as well as a 2tb Intel p3700.

If I downgrade to 24.10 the problems seem to go away?

2

Ive got the same error on my truenas setup (25.10.0), duting moderate nfs load the nfsd starts to hang and the datarate drops to a few kb.

nothing else anywhere in the logs except for that message.

did you ever find out more about that issue, or was installing an older version the only way to fix it?

3

So the problem for was two different style of nvme for the l2arc.

once we got rid of one it started to work fine.

4

what hardware are you running?