Strange messages in automount.log

[Mark Watson_2]

i keep getting the following messages in my automount.log, and i don't know what is causing them, can anyone help??

Nov 30 00:41:15 (thread 1) UNMOUNT REQUEST:
Nov 30 00:41:15 (thread 1) dev=4300000a rdev=0 indirect
Nov 30 00:41:15 (thread 1) dev=65000000 rdev=10 indirect
Nov 30 00:41:15 (thread 1) dev=6f2e6469 rdev=72656374 direct
Nov 30 00:41:15 (thread 1) UNMOUNT REPLY : status=1
Nov 30 00:41:15 (thread 4) UNMOUNT REQUEST:
Nov 30 00:41:15 (thread 4) dev=43000009 rdev=0 direct
Nov 30 00:41:15 (thread 4) dev=65000000 rdev=10 indirect
Nov 30 00:41:15 (thread 4) dev=6f2e6469 rdev=72656374 direct
Nov 30 00:41:15 (thread 4) UNMOUNT REPLY : status=1

these messages occur about every 5 minutes and appear on my console too. I'm running HPUX B.11.11 U on a 9000/800/N4000-44. All of the automounted file systems are mounted and i can see all of the lower level files.

[Yang Qin_1]

Hi, can you check syslog.log and see if you can find some related information there? It looks like automountd tried to unmount the NFS mounted filesystem for some reason (network, DNS ... problem?) and it failed to unmount.


Yang

[Peter Nikitka]

Hi,

this may be just normal messages (an expiry of a mount point is checked for the possibility of an umount) but some tracing is enabled - there was an option '-D' if I rememeber correctly.

Remove this debug/trace flag and the messages should disappear.

mfG Peter

[David Nixon]

You can extend the default timeout
for attempting umount from 5 minutes(10
in the case of Advanced AutoFS), by
using the -t option of the automount command.

You also might consider not automounting file
systems that usually remain mounted.

[Dave Olker]

Hi Mark,

A couple of things...

1) These messages indicate AutoFS is trying to unmount a filesystem and failing. The most likely cause is the filesystem is in use and cannot be unmounted. If you want to find out for sure which filesystem it is and why it isn't unmounting you can send a SIGUSR2 (kill -17) to the running automountd daemon and this will enable full-blown debug logging. That level of logging will show you which filesystem automountd is working on and why it failed - if it fails.

If you want to turn debug logging back off after seeing the log entries you're interested in you simply send another SIGUSR2 signal to the automountd process and it will toggle logging off.

2) The fact that the filesystem is attempting to unmount every 5 minutes means either you've forcibly configured a 5 minute unmount timer or you're using the older ONC 1.2-based version of AutoFS. I'd strongly recommend you obtain the newer ONC 2.3-based version of AutoFS for 11.11. This version is available as a free download from:

http://h20293.www2.hp.com/portal/swdepot/displayProductInfo.do?productNumber=ENHAUTO

This is a much newer, much better version of AutoFS. It resolves nearly every problem we've seen with the older AutoFS. In fact, the NFS lab won't even investigate problems with the older AutoFS anymore. Any customers experiencing AutoFS problems on 11.11 are directed to the ONC 2.3 version, and 999 times out of 1000 that resolves their problem.

If your shop uses AutoFS I strongly urge you to get the ONC 2.3 version. FYI - the ONC 2.3 version is what ships with HP-UX 11.23 by default.

Regards,

Dave