Question

How to reduce backup time VMs in Google Cloud

  • 6 November 2023
  • 2 replies
  • 36 views

Userlevel 3
Badge +7

Hello everyone,

 

I'm in a situation that's causing me some discomfort. I have 15 VMs located in one region and I need to replicate these VMs to another region, all within the Google Cloud environment.

To do this, I'm using 1 MediaAgent to perform the backup and replication operations. However, I'm noticing that it takes a very long time to complete the entire replication process, taking around 15 hours to complete successfully.

 

The first time running the Job, Commvault said that the Application Size was of 7.22TB. And in the last incremental, the Application Size it’s of 112.44GB.

 

Because of this long time, I’d like some tips to help reduce this time, if there’s any documentation that can help.

 

Thanks, 🙂.


2 replies

Userlevel 2
Badge +4

@antonio 

Thank you for reaching out.

Performance totally depends on Infrastructure you have and it varies from environment to environment.


You can check for below documentation where deployment planning is given. 

https://documentation.commvault.com/v11/expert/93513_deployment_planning.html

 

Please go through it and suggest if it was helpful.

If not I would suggest you to raise case with us so that we can check and narrow down issue where the actual limitation is. 

 

Userlevel 3
Badge +7

Hello @nishas, thank your reply.

 

I took a look in the documentation and the same time, I had opened a ticket to try to identify what could be causing this delay.


According to support, this is because Google Cloud uses CRC to collect change information. The suggestion was to increase the DataReaders and check that the VSA has the correct sizing for the capacity of the data, but even with the adjustment of the DataReaders' sizing and also the validation of the VSA's sizing, I see that there has been no improvement in performance.

I also created a MA in the region/location where the source VMs are, so we could maybe have a better performance with that, but doesn’t seems to have had much effect. (Now having 1 MA in the source region and 1 MA in the destination) - I haven't shared the DDB partitions or moved the DDB data from one MA to the other yet.

I currently have 1 Job that has been running up for 5 days stuck in 1 single VM.

Regards.

Reply