IBM RS/6000 SP, Problem Determination Manual

Page: 242 / 316

This soft copy for use by IBM employees only.
For further information on Amd options, refer to the man pages in
/usr/lpp/ssp/public/amd920824upl75.tar.Z. The man pages can be extracted with
the commands:
zcat amd920824upl75.tar.Z | tar -xvf - amd920824upl75/amd/amd.8
zcat amd920824upl75.tar.Z | tar -xvf - amd920824upl75/amq/amq.8
mv amd920824upl75/amd/amd.8 /usr/man/man8
mv amd920824upl75/amq/amq.8 /usr/man/man8
9.5.2 Amd Hints and Tips
1. Do not use
kill -9 <pid>
to stop the A m d daemon. Use
kill -15 <pid>
instead. Using
kill -9
can cause the Amd daemon to hang. Rebooting the
affected node is the only way to recover from the hang.
2. Before running
/etc/amd/amd_start -f
, unmount all filesystems which have
been mounted by Amd. This will ensure that the same filesystems are not
mounted multiple times, which could potentially cause problems.
root@sp21n09 / > df
Filesystem
512-blocks Free %Used Iused %Iused Mounted on
/dev/hd4 16384 6960 58% 1050 26% /
/dev/hd2 589824 11936 98% 11188 16% /usr
/dev/hd9var 8192 5440 34% 334 33% /var
/dev/hd3 24576 23320 6% 40 1% /tmp
/dev/hd1 8192 7840 5% 18 2% /home
sp21cw0:/tony 40960 39312 5% - - /a/sp21cw0/tony
sp21cw0:/tony 40960 39312 5% - - /a/sp21cw0/tony
sp21n09:(pid3472) 0 0 -1% - - /u
sp21n09:(pid3472) 0 0 -1% - - /auto
sp21cw0:/tony 40960 39312 5% - - /a/sp21cw0/tony
Figure 93. Example of Output f r o m df C o m m a n d Showing Multiple Mounts
If multiple mounts of the same filesystem have occurred (as in Figure 93),
then try to unmount all the Amd-mounted filesystems. The command
umount
/a/sp21cw0/tony
will need to be run three times in the example. Then run
/etc/amd/amd_start -f . This will clear up any problems (in most cases).
3. If filesystems which should be automounted are not, check whether the A m d
daemon is running. Also make sure that the map files in
/etc/amd/amd-maps are correctly configured. See Figure 93 as an example
of a map file. Ensure that the map files are the same on all the nodes (use
supper
to update them, if necessary). If everything looks correct, restart the
daemon (on the Control Workstation and each node) by using the
amd_start
-f
command.
222
SP PD Guide