Andrey Kiryanov wrote:
> Andreas Gellrich wrote:
>
>>Here I see a defunct process coming up ...
>>root@grid-ce2: [~] pp gma
>>root 3947 0.6 0.7 35256 29788 ? Ss 14:58 0:13 globus-gma: polling jobs
>>41753 10740 48.2 0.0 0 0 ? Z 15:26 4:26 [globus-gma] <defunct>
>
>
> Right, 41753 is a UID for ilcprd003 account which is probably in a bad
> shape. [...]
Andreas, Christoph, you probably will want to clean up that account.
You could ask the user in question if the previously submitted jobs
can all be removed, to get the CE into a healthy state again.
If so, the best would be to ban the DN to avoid interference from
WMS nodes while the cleanup is being done.
The cleanup steps then would be as follows:
--------------------------------------------------------------------
/etc/init.d/globus-gma stop
ps -u 41753 | awk '{ print $1 }' | xargs -r kill -STOP
mv ~ilcprd003/.lcgjm ~ilcprd003/.lcgjm.bad
mv ~ilcprd003/.globus ~ilcprd003/.globus.bad
mkdir /opt/globus/tmp/bad
find /opt/globus/tmp/gram_job_state/ -user 41753 -exec \
mv {} /opt/globus/tmp/bad/ \;
ps -u 41753 | awk '{ print $1 }' | xargs -r kill -9
/etc/init.d/globus-gma start
--------------------------------------------------------------------
Later the moved stuff can be deleted.
|