NFS problem possibly since PK7?

[Graham Allan]

We just had a weird issue on our ES40 file server after installing PK7. I don't know if it's related or not, but I've never seen it before, and I'm not really sure what the messages mean. I was hoping someone here might have some ideas!

To summarize, it's a 4x866MHz ES40, 16GB RAM, serving NFS to about 120 linux boxes, and samba to a bunch of Windows systems. Before PK7, we were running 5.1B PK6 with additional NFS patch OSFPATC0212700540 (fixes a readdirplus issue for linux clients).

What seemed to happen was NFS to the linux clients failed. We were able to log in to the console, but things seemed quite laggy and slow. We just rebooted to get things back to normal.

/var/adm/messages showed these:

Aug 28 14:00:48 montrose vmunix: NFS over TCP client 128.xxx.xxx.xxx not reading replies, continuing
Aug 28 14:01:17 montrose vmunix: NFS over TCP client 128.xxx.xxx.yyy not reading replies, continuing
Aug 28 14:02:34 montrose vmunix: NFS over TCP client 128.xxx.xxx.xxx not reading replies, continuing
Aug 28 14:06:26 montrose vmunix: NFS over TCP client 128.xxx.xxx.yyy not reading replies, continuing

and after a long pause, when we suspected the system may have hung completely:

Aug 28 14:17:17 montrose vmunix: Wait times greater than threshold (10) seconds
Aug 28 14:17:17 montrose vmunix:
Aug 28 14:17:17 montrose vmunix: Wait times greater than threshold (10) seconds
Aug 28 14:17:17 montrose last message repeated 25 times

Does anyone have any idea what these messages imply?

Thanks,

Graham