MWS Failover Clustering


Iam trying to implement the MWS 7.1.2 failover hardware clustering at OS level.

a) Installed MWS on Server A (Active).
b) Created the startup, shutdown scripts for veritas VCS software to unmount and mount to Server B (Passive).
c) Virtual IP has been created and mapped in the /etc/hosts.
d) from MWS GUI (Administration → MyWebMethods ->Cluster Settings), changed the front end url to Virtual IP(Logical). Also changed the HostName to Virtual IP.

Now, I can able to logon the MWS on Server A using the front end url provided on step d.

But when I try to shutdown the server using ./ -s default stop &

The follwoing error has been thrown.

JAVA_HOME: /Portal/jvm/sol150
SERVER_HOME: /Portal/MWS/server/default
Stopping MWS…
loading configuration from MWS database…
shutting down:
2008-12-08 16:33:57 EST (Framework : INFO) [RID:8] - Processing request [1kz2zv7dakcgl:null] [URL][/URL] (GET)
2008-12-08 16:33:57 EST (Framework : WARN) [RID:8] - Checking Local Control addr:
invalid stream header invalid stream header
at com.webMethods.rtl.util.ObjectUtil$ClassLoaderObjectInputStream.(
at com.webMethods.rtl.util.ObjectUtil.load(
at com.webMethods.rtl.util.ObjectUtil.load(
at com.webMethods.portal.system.SystemControlUtil.connect(SystemControl

Even I tried using -n option to specify the node name of the cluster with no luck.

Checked the cluster.xml (need to get it from DB from verison 7.1.2) and it looked same as in the “Advanced or Clustered Configuration” MWS clustering page.

Not sure where I went wrong. Any help is appreciated.

I know this is not the support mechanism as per the documentation. But was there a resolution to getting “./ stop” to work correctly within the cluster / across the cluster nodes.