Earlier today I had a request to nfs mount a couple of partitons. When I went to the client, I found the entries in /etc/fstab already:
Code:
store1:/shared/radsm_npp /radsm_npp nfs rw,bg,hard,intr,vers=3,proto=tcp,timeo=600,retrans=2,rsize=32768,wsize=32768,tcp 0 0
store1:/shared/radsm_data_repos /radsm_data_repos nfs rw,bg,hard,intr,vers=3,proto=tcp,timeo=600,retrans=2,rsize=32768,wsize=32768,tcp 0 0
On the server side I could see the request being made, but the mount wouldn't take place. Someone said the had ran across this problem recently and used udp instead of tcp. I changed ONLY the 2nd tcp entry and it worked:
Code:
store1:/shared/radsm_npp /radsm_npp nfs rw,bg,hard,intr,vers=3,proto=tcp,timeo=600,retrans=2,rsize=32768,wsize=32768,udp 0 0
store1:/shared/radsm_data_repos /radsm_data_repos nfs rw,bg,hard,intr,vers=3,proto=tcp,timeo=600,retrans=2,rsize=32768,wsize=32768,udp 0 0
The man page talks about tcp and udp, but why would there be two references to the protocol on the same line. And why does changing just one work? I frankly didn't notice the first one, and it was by accident it worked. Should both entries be changed?