Skip to main content

 

A customer is protecting Oracle running in a failover cluster between Solaris Local Zones. Customer is currently using TSM and want to switch to Commvault. 

In TSM the Oracle databases are protected with the Oracle Agent running in the Solaris Global Zone only, there is no agent in the Clustered Local Zones. This means that no matter in which Local Zone the Ora DB is running the backup and restores will work without setting this up as cluster client. 

The question is if this is possible in a Commvault setup or if we are required to install the agent in each Local Zone and then configure the cluster client in a traditional way.  

Appreciate if anyone has experience of a similar setup has any input if this is at all possible and if so any  pros & cons.

Regards,

/Patrik 

Hi @Patrik 

Thank you for the question.

Checking Oracle: System Requirements:

Solaris

In a UNIX virtualization environment using Solaris, you can back up and restore data in both global and non-global zones.

Then again, there is a little more suggestion from the linked page UNIX Virtualization:

Oracle

It is recommended to install the software on the global zone to protect non-changing or static data on non-global zones. If the data is dynamic or contains application data, install the software on the non-global zone.

Thanks,

Stuart

 


Thanks Stuart, 

I saw that but wanted to check if it would be possible to accomplish a similar function with Commvault as with the current setup with TSM. Customer thinks the recommended way introduce more complexity but from a Commvault perspective maybe not as it is a standard failover cluster. 

I guess it will come down to testing and verifying if no one has any experiences of a similar apporach. 

Regards,

/Patrik  

 


Testing has been done. 

We did not setup standard cluster client with two solaris local zones (cluster nodes) clients. 

Instead we installed the client in the cluster resource group itself so the entire cv installation will follow the resource group on cluster failover.

We had to link the /etc/galaxy-registry folder to a cluster resource drive and mount the /opt/commvault as well as the commvualt/Log_Files folders from the cluster resource drives. 

We did backup and restore tests, failed over the cluster resource group and continued operations successfully. 

In this way we reduce the number of clients by half (if not including the cluster client in normal setup). Imagine the impact on the oi licensing, even though that did’nt matter here at the end as customer will use capacity license, but it make the setup cleaner. 

/Patrik   


Thanks for sharing!   Let me know if you need further help here, or if we can consider this resolved.


Reply