Automounter is hung up

[Greg OBarr]

HP-UX 11.0 automounter has caused a problem and I can't "ls -l /" - it's complaining, saying "NFS server (pid585@/net) not responding still trying". I have tried killing all the NFS daemons, rpc.lockd, etc using /sbin/init.d/nfs.core,client, and server. I have seen this before and fixed by rebooting. Anyone know how to fix without rebooting?

[Brian Hackley · Accepted solution]

Greg,

Does ps -ef |grep 585 show process ID 585 still running? If not, you have to reboot.

If there is an automount PID 585 running, then
send a SIGUSR2 (kill -17 585) to see if anything shows up in /var/adm/automount.log
If that does not wake up the automounter, most like a reboot is in order.

As an FYI, the automount process "fakes out" the kernel into its process as managing the automount points, emulating an NFS Server. In this case, it is the /net indirect "-hosts" map. If the process "goes away" without the kernel being told to close out the Virtual Filesystem mountpoint, then any process that tries to access that mountpoint will block (sleep) forever, since the process is no longer there to answer the request.

Hope this helps,

-> Brian Hackley

[Greg OBarr]

Yes, the automounter has been restarted, so the PID is no longer 585. This answers my question, although it's not what I wanted to hear. I wonder what the effect would be if I could get automount to start again at PID 585?.... just out of curiosity.