nfslock not running at startup

[Andrew Kaplan]

We have a group of Red Hat 7.3 servers running a clustering application. One peculiarity that occurs is when a system is brought down and subsequently rebooted, the nfslock daemon does not automatically activate on startup. I am
forced to manually start the daemon. Has anyone had this problem occur on their systems, and if so what was the fix? Thanks.

[Alvaro Miranda Aguilera]

Hi,

Please include some more info.

chkconfig --list

You need for nsf services, poprtmap, nfs and nfslock.

This machine is a nfs export server? or a nfs client?.

Alvaro.

[Andrew Kaplan]

Hi there --

I've included a printout of the chkconfig command for your perusal. Thanks again for your reply.

[Alvaro Miranda Aguilera]

Hi Andrew,

You don't answerme when I ask about if your machine is a Server or a Client.

I'm supossing it is a client.

Do a chkconfig portmap on

and Try, If does'n work, try this: Disable the nfslock and try to mount the nfs folder anyway.

I'm stuck with your problem because i'm guessing many things.

By other hand, the bests practices for your needing, is enable and configure the autofs, so your linux will mount and dismount the folder when is necesary.

Alvaro.

[Serviceguard for Linux]

It's my understanding that for most versions of NFS on Linux, the NFS lock will not survive a failover to another machine.

Some clusters that support it have special code to handle that case.

[Michael Williams_6]

Andrew, one possibility I can think of is if you have clustered you /var/lib/nfs directory as you're supposed to.

When the system comes up, is the /var/lib/nfs directory actually available at the time, or is it still on the other node?

I am pressuming here that the cluster is the root cause of the problem and that you have a shared filesystem using drbd or something. Can you supply more information about how you've built the cluster?

[Michael Williams_6]

Ignore this message, forgot to tick the sent me an e-mail when someone replies option!