Automount problem

[Jerome D'silva]

Hi,
I have a 9000 L class server runing HP UX 11.0.
Sometimes automount doesnot function and I can see this error in the syslog "NFs server (pid921@/net) not responding still trying" .
When I do /net , I get this same error . This server has been running from past two years without any problem and recently this problem started .
I restarted the system , but after sometime the same errors were appearing .This problem is not constant , but appears once in two to three hours and after sometime the automount functions properly. The syslog throws this message "NFS server (pid921@/net) ok" .

Do I need to install any patch for it ?. Please reply to this mail at the earliest .

Thanks
Jerome

[Vijeesh CTK]

hi

check in your /etc/auto_master file. hash the entry in that file(/net -hosts -nosuid,soft)
and then restart nfs.client

/sbin/init.d/nfs.client stop
/sbin/init.d/nfs.client start

ur problem will get solved

CTK

[Michael Tully]

Hi Jerome,

Here is the latest patches for NFS and automount that I could find. These do not take into consideration each and every dependency.

PHKL_23409
PHKL_24082
PHKL_25237
PHNE_20627
PHCO_21187
PHCO_21901
PHKL_25475
PHNE_20371

Were any changes done on the client side?

Michael

[Hartmut Lang]

I think your problem can also be caused by networking problems. If your network get congested some time, the NFS-client cannot reach the server: "still trying". After network traffic get normal the connections will be made.
Or your server is very busy during this time, so it cannot answer the mount-requests as expected.
Going to latest patches (or latest Quality Pack) will be a good thing anyhow.

Hartmut

[PIYUSH D. PATEL]

Hi,

If you have not changed any of the automounter files then the problem is with some network connection only.

Since your server was working fine earlier and if you havent updated anything then patches may not be a problem.

Check for the network connectivity.

Piyush

[Brian Hackley]

Jerome,

The automount process ID 921 is not responding to the file system request to access it. It is possible that the automount is hung or otherwise tied up. You can put the automount into debug mode with a kill -17 921 (or whatever the PID is). The ISGUSR2 signal is an on/off switch for debug mode. The log file goes to /var/adm/automount.log. Its possible you might find some cryptic messages from automount that might expose other reasons for the automount hang up.
As noted in some of of the prior responses on this thread, you might want to patch up the automount and NFS in general, especially if you have not done so in the past year.
I also concur, that you want to comment out the /net -hosts entry in /etc/auto_master if you don't plan to use the /net map. Hierarchical -hosts maps cause the automount to query the NFS server for all of its exports, and then to mount them all, even if they are not needed. Then, every 5 minutes, the automount attempts to umount that /net/ structure. If any umount attempt gets EBUSY (e.g. mount point in use), then automount has to re-mount all the mount points in that hosts /net entry. Its complicated and alot of overhead, and sometimes, a non-responsive or bad network connection to that server can cause the automount process to be tied up and not responding. Hence the messages you see.

Hope this helps,

-> Brian Hackley

[Shannon Petry]

This could also be a load problem. If the /net/ directory is getting lots of use, then try bumping up the biod number from the default 4 to 16. This is in the file /etc/rc.config.d/nfsconf.

Regards,
Shannon