Hi guys, I am struggling to find a balance between stream and datastore limits for our environment which has many datastores and clusters on a shared flash array (using SAN transport). The VSA client also contains numerous remote sites all backing over over NBD, all of which funnel to the same MA/access node. The MA has plenty of resources but is fast enough to really hammer the array.
https://documentation.commvault.com/11.21/expert/31115_performance_tuning_for_backups.html is what I’m referencing but I see no clear cut way to throttle TOTAL datastore streams hitting the SAN but not limit all the various NBD-based remote subclients. Meaning - I can’t limit the streams of the MA without impacting the run times of all those NBD clients because they take hours to complete in some cases. Basically what I’m looking for is a way to have the proxy/access node tell “hey, all these subclients and datastores reside on the same array, so limit total streams for all those combined SAN streams to say 10” without slowing down the non-SAN based transport jobs. Is this possible?
I’d prefer to not have to create lots of plans to split things out 30 different ways but this seems like the only way (like I used to do pre-plan days).
thanks