I have a two-node cluster, version Compaq Tru64 UNIX V5.1 (Rev. 732), which is a NFS server for a large farm of Linux clients. I have some client in the same subnet as the cluster, 16.11.0/24 and can successful copy large files, 1 minute to copy. Then I have clients in subnet 16.11.1/24 which receive an
'Input/output error' every time. A work around to this problem was to reduce the wsize to 1024, now I can copies these large files but it take 10 minutes or more to copy.
It looks like packets are being lost between the NFS client and the router. I have some
tcpdump outputs that show this.
The network looks simething like this:
linux -100M- cisco 5505 -100M-+
Tru64 -100M - cisco 5900-Fddi-+
any help would be appreciated.