NFS Services Administrator's Guide (762805-001, March 2014)

May 13 18:45:09 t5 args_temp: hpnfs127, , 0x3004060, 0, 0, 0, 0,0, 0, 0, 0, hpnfs127:/tmp
May 13 18:45:09 t5 mount hpnfs127:/tmp dev=44000004 rdev=0 OK
May 13 18:45:09 t5 MOUNT REPLY: status=0, AUTOFS_DONE
Unmount Event Tracing Output
The general format of an unmount event trace is:
UNMOUNT REQUEST:<time stamp>
<other tracing>
...
<other tracing>
UNMOUNT REPLY: <status>=unmount status
The unmount status in the unmount reply contains 0 if the unmount is successful; it has a
non-zero value when the unmount is not successful.
The following is an example of a typical unmount trace event:
May 13 18:46:27 t1 UNMOUNT REQUEST: Tue May 13 18:46:27 2003
May 13 18:46:27 t1 dev=44000004 rdev=0 direct
May 13 18:46:27 t1 ping: hpnfs127 request vers=3 min=2
May 13 18:46:27 t1 pingnfs OK: nfs version=3
May 13 18:46:27 t1 nfsunmount: umount /n2ktmp_8264/nfs127/tmp
May 13 18:46:27 t1 Port numbers are 937, 937
May 13 18:46:27 t1 Port match
May 13 18:46:27 t1 nfsunmount: umount /n2ktmp_8264/nfs127/tmp OK
May 13 18:46:27 t1 unmount /n2ktmp_8264/nfs127/tmp OK
May 13 18:46:27 t1 UNMOUNT REPLY: status=0
NOTE: Autofskd initiates kernel unmount thread every 120 seconds to check the mount points
that are ready to be unmounted. The unmount operation of autofs mount points are controlled
through the parameter AUTOMOUNT –t <timeout value> and the default timeout value is 10
minutes.
Consider a scenario, where the kernel unmount thread gets triggered just before the 10th minute
mount point timeout period and the mount point is not yet ready to be unmounted. In this case, the
mount point needs to wait for another 120 seconds until autofskd kernel unmount thread initiates
unmount operation on the mount point. Thus, it takes <user specified unmount timeout> + <autofskd
kernel unmount thread timeout>to unmount the file system which in this case comes to about
12minutes (10+ 2 minutes).
When the unmount-thread runs, it unmounts only one level of mountpoints at a time. This means
that on a nested mount, there is another 2 minutes for each level of unmount operation. E.g. a
“/net/host/var/opt/ignite/clients” path, with 3 filesystems under it (“/net/host/var”,
“/net/host/var/opt”, “/net/host/var/opt/ignite/clients”) will take 10+2+2+2=16 minutes to
unmount.
A “/sbin/initd/autofs stop” does not remove any /net indirect mount points in kernel.
AutoFS configuration in build environment
AutoFS automatically mounts and unmounts file systems after a period of inactivity. When the build
activity is taking place, the linker caches the object file's device id. If the build activity takes more
time than AutoFS inactivity time, AutoFS unmounts and remounts the NFS file system. This unmount
and remount of the NFS filesystem results in a change to the NFS Filesystem device ID. This can
result in build failure when the linker re-validates the cached NFS Filesystem device ID.
The workaround for this issue is to increase the AutoFS unmount time. Refer to the “AutoFS
configuration changes” (page 58) section for information on changing the AutoFS unmount time.
Troubleshooting AutoFS 79