Hi all,
we have a problem with the configuration of a linux cluster based on RH
ES3 WS. The cluster contains 230+ nodes; these nodes are sharing
/usr/local/bin over NFS due to shared scripts for health checking.
Now, if we encounter NFS problems, the login process is hanging non-stop
until the NFS connection is back again.
After checking all startup scripts and $PATH definitions we found out,
that the /bin/login binary itself is pre-defining the $PATH hardwired
with /usr/local/bin in the beginning.
Unfortunately we cannot change the /usr/local/bin down to the local disk
again, but need a solution to be able to login even or especially with
NFS problems to start investigation.
Do you have any nice idea, how to get rid of this problem ? Changing the
binary and rebuild the package is one, but we don't like it because of
missing update possibility after it ...
Thnx in advance
Cheers
Alex
--
=======================================================
Alexander Spanke System Analyst
NEC High Performance Computing Europe GmbH
Prinzenallee 11 D-40549 Duesseldorf, Germany
Tel: +49 211 5369 146 aspanke@xxxxxxxxxxxx
Fax: +49 211 5369 199 http://www.hpce.nec.com
=======================================================
--
Linux-cluster@xxxxxxxxxx
http://www.redhat.com/mailman/listinfo/linux-cluster
