[Linux-cluster] Ping periodically failed and ip.sh under /usr/share/cluster

[KC LO]

Dear support,

We have setup a Redhat cluster with floating IP.  The whole cluster is
running healthy but occasionally it occurs error messages on "Failed to
ping <IP>" and cause node fail-over.  It happend from nodea to nodeb and
then nodeb to nodea.  The network is running healthy.  I have searched over
Internet and it may relate with the ping program with "segmentation
fault".  Some workaround is to modify the program of
/usr/share/cluster/ip.sh script.

I have modified the script(/usr/share/cluster/ip.sh) so that it always
return true on ping.  It can solve the problem some period of time(around 1
month) but the problem occurred again in past few weeks.  I have inserted
some code(like insert date/time info to a log file) to
/usr/share/cluster/ip.sh to  check for any activities of the script.  The
strange is that nothing has been output from /usr/share/cluster/ip.sh but
the cluster still report "Failed to ping <IP>".

For this ping check activity, which program/script is using so that I can
modify?
Can I disable the ping check?

Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20120227/0f899459/attachment.htm>