Regards,
Today we have had hell trying to figure out why Sun (Solaris 2.5 and 2.5.1)
NFS clients do not want to talk to our Alpha NFS server(s). SunOS 4.1.3
clients work OK. The servers are configured in a ASE (1.4A) cluster running
4.0B (patchkit #5). They are also connected to a FDDI ring (fta controllers)
and the Sun clients that have trouble getting NFS traffic through are all
sitting on a switched ethernet (both fast and regular). Between the servers
and the clients there are Cisco switches.The symptoms are the usual "server
not responding" messages on the clients. On the servers, however, there are
messages saying
Mar 17 16:01:53 server vmunix: svcktcp_reply: mbuf_send returned 32
Mar 17 16:01:53 server vmunix: rfs_dispatch: sendreply failed
netstat -m output seems normal, we have tried powercycling both servers in
the ASE cluster but no success. We have also tried rebooting SUN clients
with no luck.
The peculiar in the situation is that the errors are spurious, there seem to
be no consistency. One moment all is OK, the next everything hangs on the
SUN clients and the above messages appear on the server(s). Also, all other
client architectures work OK, both SGI, IBM, and other Alpha clients work
all the time (and, as I said, SunOS 4.1.3 clients). I do not know if this is
relevant, but all switches got their software upgraded a couple of weeks
ago, but things have worked well since then. The NFS traffic is also mixed
nfs2/nfs3 that work OK on the IBM/SGI clients.
Any ideas/suggestions are highly appreciated !!!!!
--
******************************************************************
* Knut Hellebų | DAMN GOOD COFFEE !! *
* Norsk Hydro a.s | (and hot too) *
* Phone: +47 55 996870, Fax: +47 55 996495 | *
* Cellular Phone: +47 93092402 | *
* E-mail: Knut.Hellebo_at_nho.hydro.com | Dale Cooper, FBI *
******************************************************************
Received on Tue Mar 17 1998 - 16:41:37 NZST