Tags based Intellisnap

Question

We have newly provisioned HSX platform and we are doing VSA backups on Vmware 8

We are utilizing Intellisnap

At the moment we have one job using PRD tag and another job doing non-prd and UAT tags

These are HSX nodes (11) and over 1000 VMs

Although the backups are running the backup copies seem to be taking a long time (NBD; we will migrate to SAN mode)

Also, we have tag based exclusions; so if a VM is both in excluded tag(RDBMS should not be snapshotted) and included tag (Some RDBMS should be which dont have agent based backup) the VMs does not get backed up; we are currently manually adding those vms to backup and that seem to work.

Is there a better way to use tags say we will have to add specific VMs to configuration

In our old Commvault environment we have VM data-store based jobs (IntelliSnap again) but these had virtual proxies and because each ds backup triggered at a different time and separate virtual proxies(hot-add); it seemed more efficient in terms of backup copies.

Now the proxies are physical and we are running one job; it does not feel right; should we be breaking up the jobs into multiple jobs staggered across night; this is to make sure one job failing say if a vm is across 2 datastores does not fail all the vms; is there a way to achive that or is it all changed in Commvault and in the background now jobs are per vm based.

Also looking at the backup copies it seem to be utilizing only 2 nodes although we have multiple nodes ; how do we ensure that we have maximize and utilize other nodes as well

Bronco · Answer

Hello Rajeev,current Commvault hsx backup setup and identified a few areas that can improve performance, reliability, and overall management. Today, the environment uses large backup jobs grouped by environment (Production and Non-Production), but this creates a larger impact if a failure occurs and limits how efficiently the system can use all available HSX nodes. We recommend splitting these into smaller, logically grouped backup jobs (for example by datastore, application, or business unit) and staggering schedules. This will improve parallel processing, reduce the impact of failures, and make troubleshooting much easier.For VM selection, the current tag-based approach is working as designed if a VM matches both an inclusion and exclusion tag, the exclusion wins. The current workaround of manually adding important VMs into the backup content is the correct and supported approach. Longer term, refining the VMware tagging strategy may reduce the need for manual exceptions and simplify administration.We also observed that backup copy operations are currently utilizing only a small number of the available HSX nodes, which limits throughput. By increasing the number of concurrent jobs, adjusting stream settings, and ensuring all HSX nodes are configured as access nodes, the environment can better distribute workloads across the cluster. In addition, moving from NBD transport mode to SAN mode is strongly recommended, as SAN provides significantly faster data movement and better scalability for large VMware environments.Regards,Siddharth Gandhi

Tags based Intellisnap

1 reply

Readiverse Academy Certifications Tier

Readiverse Academy Certifications Tier

Sign up

Login to the community

Scanning file for viruses.

This file cannot be downloaded