Skip to main content

Hi,

we have a unix (ubuntu linux) cluster and installed commvault on all three nodes. After that I created a “unix cluster” with there three nodes with the virtual cluster IP.

I could do a backup for the cluster filesystem, but after a failover of the cluster filesystem to another cluster node, the backup failed. It still tries the backup on the previous node. So I think the local commvault agents didn’t notice the node change. Is there any chance to trigger the node change to the commvault agent?

As a workaround, I restarted the commvault agent locally, but a solution for running services would be better.

Thanks,
Janko

Hi @Jankete 

 

Right click the cluster client and click Advanced. In the bottom left corner on the Cluster Group Configuration pane, tick off the “Force Sync configuration on remote nodes”. Click Ok.. 

 

/Rubeck

   

 

Hi, thanks for that. But what I’m looking for is some kind of an automatic way. When a cluster failover happens at 3am on Saturday, the CommVault agent should recognize this without a manual action. Otherwise a cluster object on doesn’t make much sense.

@Jankete , I believe I have what you are looking for here:

https://documentation.commvault.com/11.24/expert/4990_configuring_cluster_group_clients_for_non_veritas_clusters_unix.html

The Force Sync configuration on remote nodes option is for applying updates after the fact:

https://documentation.commvault.com/11.24/expert/5042_troubleshooting_for_cluster_configuration.html

Hi,

We tried the cvclusternotify command, but we’re not sure how to use it? Just use the “-start” on the active node after a failover? If it war an unplanned failover, the “-shutdown” can’t be triggered on the failed node.

Hi @Jankete ,

 

You need to add cvclusternotify -shutdown script in the cluster shutdown script & cvclusternotify -start script in the startup script of the cluster.

 

When a failover happens first it runs the shutdown script on the current node and start script on the node which will take over 

 

Link → https://documentation.commvault.com/11.24/expert/4990_configuring_cluster_group_clients_for_non_veritas_clusters_unix.html

Usage:

cvclusternotify -inst InstanceName -cn ClientName m-start | -shutdown]

Where:

  • cvclusternotify is the script that notifies the Commvault software about cluster failovers.

  • -inst specifies the name of the Commvault instance on which you want to run the script. If you have a single instance, specify Instance001.

  • -cn is the name of the cluster group client.

Example:

For a two-node cluster, if the cluster group client name is "ClusterGroup1" and the application instance is "Instance001", run the following command:

  • To shutdown:

    cvclusternotify -inst Instance001 -cn "ClusterGroup1" -shutdown

  • To start up:

    cvclusternotify -inst Instance001 -cn "ClusterGroup1" -start

Thanks for that @Gowri Shankar . What about an unplanned failover? If the primary node simply crashes so that the “-shutdown” cant be sent anymore, would this be a problem?

@Jankete  we completely rely clustering software to run the shutdown / startup scripts however in case if they are not executed by any chance then you can perform a force resync from the commvault console.

 

From the commvault console → right click on the cluster pseudo client  → properties → advanced options → cluster configuration and select force resync and click on Ok

 

This will query the clustering software and make the active node in Commvault.

 

Terms & ConditionsCookie settings