Service Guard doesn't switch over

[Ionut Grigorescu_2]

Hello,

I have a HP-UX 11.11 Cluster with 2 nodes . On the so called system node are running 7 packages, one of them is monitoring the NFS server, one monitors system related functionality, the other are application related . On the other node (database) is running again a package which monitors system related functionality and another one which monitors the Oracle database. I have to test the high availability of the cluster - first I have unpluged all the LAN cables from the system node. The database node performed a reboot (in order to preserve data integrity on the nfs mounts) then tried to restart the cluster on the database node- it wasn't successfull, in the syslog I have seen entries telling - that nfs package is unknown, after few seconds try again to restart it , failed again and so on. Normally if the nfs server is not available it should be started on the present node. Can I configure a timeout somewhere so that the nfs server will be locally started if not available for a number of minutes or so?

thank you in advance,
Ionut

[Steven E. Protter]

Shalom,

Probable Cause: Cluster configuration files not the same on both systems or more likely the package control scripts are different and do not reside on both nodes.

The cluster configuration itself could be wrong and you may simply need to do further testing and get it right.

SEP

[LinuxNotMS]

Make sure you have the secondary node actually listed in the package. If it isn't you'll have to change that and re-apply that package configuration. You may also be able to get a better idea of what's going on (or what isn't) by looking at the output of cmviewcl -v.

Perhaps readers of the forum could be of more help to you if you post that output from cmviewcl -v. Then we could help make more intelligent recommendations regarding package dependancies, etc.