Skip to main content
commvault.com commvault.com
Question

mix of Fast and slower disk in one pool on a Spill and Fill?

  • December 12, 2024
  • 1 reply
  • 54 views
L
Novice

Looking for advice or documentation to support changes.  Running 11.28.126

I need to add slower (cheaper) disk storage to my Commvault backups.   I have a DISA Cloud virtual environment and it offers 2 speeds of primary disk.   Tire 1 High Performance Tire 2 Moderate performance

Tier 1 -~ all Flash/SSD disk drives

Tier 2 -~ 7200 RPM Serial Attached SCSI (SAS) disk drives

My Goal is to add the slower disk to hold about +60% of my backups.  Then the Faster disk can be used for the daily backups and quick restores and the slower disk can be used for long term storage (backups older than 7 days) and old archive restores

 

I have a single, all in 1 Commserv/Media agent… server.   It currently has 16TB of local storage Disk attached to it in 2TB disk increments (8 storage disks). Only one storage pool.  All setup as Spill and Fill.

Option A: I see in current documentation that I should use “dedicated storage pools for high-performance disks and slower disks”. 

Option B: But I also see documentation that says if I have a “mix of Fast and slower disk in one pool on a Spill and Fill, the Fast disk will be used for more current data by default” and the older data will be maintained on the slower disk.

What are the pros and cons of these options.

1 reply

Damian Andre
Vaulter
Forum|alt.badge.img+24
  • Damian AndreVaulter
  • Vaulter
  • December 14, 2024

There is no special logic to spill and fill - it will round robin data between available paths. I’m not sure where you saw the bit about maintaining older data on slower disks - I don’t feel like any special logic like that exists, especially if you are using de-duplication.

Option A is your best bet, land the data to a spool copy (no retention). These types of copies exist only to store the data until it is copied to somewhere else. Upon successful copy the data is removed from the source.

A useful setting in this case might be “ Pick data from running backup jobs “ this means you can immediately start copying data from the spool copy even before the job has completed. Its a setting on the copy policies.

 

The only thing to lookout for is if the primary spool copy will be able to hold a full baseline of all data for your full backups - i.e it wont run out of space during your weekly full jobs.

 

Terms & ConditionsCookie settingsAccessibility statement