Server DNS hang, NFS share from Netapp machine won't work

[Diyana]

It appears that the system experienced a hung NFS mount which, a reboot alone cannot be solved. It was identified earlier that a DNS server is down which has caused the hung NFS mount.

[Steven Schweda]

> [...] the system [...]

Not a very detailed description of anything.

> [...] It was identified earlier [...]

Is this posing related to your other one?

http://h30499.www3.hp.com/t5/General/linux-server-was-hung-and-causing-a-web-service-outage/m-p/4776422#M19050

If so, then why start a new thread? If it's
a new problem, then what do you want?

[Diyana]

This is not related http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1476707&admit=109447626+1303366351107+28353475.

The problem is :-

1. We have reboot client server cause the nfs hung
2. After rebooted, we have found that, the automount point is not exported from the netapp server server to our client server.
3. Netapp admin told that primary DNS server has been down for while.
4. After DNS server is up, we able to see all the mount point.

My question :-

1. How to fix this issue cause it's likely to happen again.

[Vladimir Fabecic]

Put your server IP in NetApp's /etc/hosts file.
ssh to NetApp filer, login as root and
# wrfile -a /etc/hosts x.x.x.x yourservername

[Hiren N Dave · Accepted solution]

Edit fstab and use "soft" option to mount NFS file systems to prevent from hanging while NFS server is unavailable.

When an NFS client attempts to mount a resource using the hard option, which is default, the client keeps trying until either it succeeds or is interrupted with Ctrl+c or the kill command if "intr" option is also used.

If the NFS server goes down, processes using the mounted resource hang until the server comes back up.

Use "soft" to prevent NFS clients from being hung if the NFS server goes down. With this option, if an NFS client attemps to mount a resource for "retrans" times unsuccessfully, it displays an error message. Use timeo=n and retrans=n option with this to limit number of retry attemps.