Simpler Navigation for Servers and Operating Systems - Please Update Your Bookmarks
Completed: a much simpler Servers and Operating Systems section of the Community. We combined many of the older boards, so you won't have to click through so many levels to get at the information you need. Check the consolidated boards here as many sub-forums are now single boards.
If you have bookmarked forums or discussion boards in Servers and Operating Systems, we suggest you check and update them as needed.
General
cancel
Showing results for 
Search instead for 
Did you mean: 

Problem in NFS file system mounting

sesemi
Occasional Visitor

Problem in NFS file system mounting

I have a NFS server which is exported a directory. There are 5 NFS client machines which they mount that directory too. The service used to work fine before. Now we encounter a strange problem. The mounted NFS filesystem on the clients loses its permission suddenly. The problem is that the directory on the clients loses their permissions occasionally. Although it is still mounted I cannot either change directory to the folder. I get permission denied error and the output of ls -l is as follow:
?------------ ? ? ? ? ? SharedFolder
Also, before the problem occurrence all the created files owner were root:root, now the ownership of newly created files are nobody:nobody. The Operating system of both clients and server are RHEL4-64bit.
Remounting the partition solves the problem temporarily but it happens again within an hour. The result of NFS debug process are as follow:
Client
Mar 9 12:17:04 app4 kernel: NFS: revalidating (0:13/2)
Mar 9 12:17:04 app4 kernel: NFS call getattr
Mar 9 12:17:04 app4 kernel: NFS reply getattr
Mar 9 12:17:04 app4 kernel: nfs_revalidate_inode: (0:13/2) getattr failed, error=-13

Server
Mar 9 12:14:18 app9 kernel: svc: tcp_recv 00000100782f9a80 data 1 conn 0 close 0
Mar 9 12:14:18 app9 kernel: svc: tcp_recv 00000100782f9a80 data 1 conn 0 close 0
Mar 9 12:14:18 app9 kernel: svc: socket 00000100782f9a80 recvfrom(00000100782f9b28, 4) = -11
Mar 9 12:14:18 app9 kernel: RPC: TCP recvfrom got EAGAIN
Mar 9 12:14:18 app9 kernel: svc: got len=-11
Mar 9 12:14:18 app9 kernel: svc: server 000001007e177800 waiting for data (to = 3600000)
Mar 9 12:14:18 app9 kernel: svc: server 000001007de41c00 waiting for data (to = 3600000
6 REPLIES
Modris Bremze
Esteemed Contributor

Re: Problem in NFS file system mounting

What are the /etc/exports options on server and mount options (/etc/fstab?) on clients? What does nfsstat show?
Modris Bremze
Esteemed Contributor

Re: Problem in NFS file system mounting

What are the /etc/exports options on server and mount options (/etc/fstab?) on clients? What does `nfsstat` show?
sesemi
Occasional Visitor

Re: Problem in NFS file system mounting

the etc/fstab configuration for this mount point is as follow:

x.x.x.x:/SharedFolder /SharedFolder nfs remount,rw,hard,nointr,tcp,vers=3,timeo=300,rsize= 32768,wsize=32768,actimeo=0 0 0



The output of mount command when the problem happens:
x.x.x.x:/SharedFolder /SharedFolder type nfs (rw,hard,nointr,nfsvers=3,tcp,timeo=300,rsize=3276 8,wsize=32768,actimeo=0,addr=x.x.x.x)

This the content of /etc/exports
/SharedFolder app*.domainname(rw)
/SharedFolder x.x.x*(rw)
Kapil Jha
Honored Contributor

Re: Problem in NFS file system mounting

You would like to check on crontabs and other scripts which could possibly cause the permission to be changed.
Now since earlier permission used to be root:root that mean only root can change the permissions so you would like to see if any command/script.
You may also would like to check if there is any sequence for permission change.

BR,
Kapil+

I am in this small bowl, I wane see the real world......
sesemi
Occasional Visitor

Re: Problem in NFS file system mounting

Nothing is set in crontab.
Modris Bremze
Esteemed Contributor

Re: Problem in NFS file system mounting

You could try mounting it with default options (defaults) and see if the problem persists. You also mentioned, thtat there are multiple clients accessing the share - maybe try with only one client at a time. It would also be useful to ensure, that all relevant patches are applied.