Hi all,
I've got an EMC NFS head and a bunch of FC2 machines mounting it
(2.6.9-1.6_FC2smp, have also tried latest 2.4 with same results). A few
mail boxes which don't have any problems and four web servers which hang
constantly. The build is the same but the web boxes need locking (for
cgi's etc).
Pretty much twice a day one of the machines goes pop.
Here is what's in 'dmesg' :
nfs_statfs: statfs error = 512
nfs_statfs: statfs error = 512
nfs: server nfs not responding, still trying
nfs: server nfs not responding, still trying
nfs: server nfs not responding, still trying
nfs: server nfs not responding, still trying
nfs: server nfs not responding, still trying
nfs: server nfs not responding, still trying
nfs: server nfs not responding, still trying
nfs: server nfs not responding, still trying
nfs: server nfs not responding, still trying
nfs: server nfs not responding, still trying
<and so on>
Here is a snippet from an ethereal at the time.
NFS [RPC retransmission of
#51]V3 WRITE Call (Reply In 175), FH:0xdc1a4c96 Offset:122880 Len:8192
UNSTABLE[Unreassembled Packet]
Now as far as I can tell the NFS server doesn't die or reboot (as the
512 error code would indicate) and the other machines keep working.
Anybody see this sort of thing before or have any good nfs
troubleshooting tips from the client side? We are pretty much locked out
of the server so what we can check is limited.
Cheers.
