Highlighted
1 Copper

XtremIO Snapshots (XVC) Inquiries

1. Is there a way to limit storage capacity consumed by a snapshot so as to avoid fully consuming the storage pool? 2. Is there a way to control snapshot behavior if it is about to consume all the storage capacity available to prevent from fully utilizing the entire pool? Thanks in advance!
0 Kudos
Reply
1 Reply
Highlighted
2 Bronze

Hello CarlCamper,

 

XtremIO snapshots in of themselves do not consume any space on the disk in the DAE. At the point the snapshot is created every bit of data from the source volume is already committed to disk in the DAE. As such the data deduplicates 100%. Now if the source volume and the snapshot are both mapped to host, any changes to data or new data written may consume more space on the array. I say may as there is still potential for deduplication.

Another point of snapshots, from array perspective volumes and snapshots are all treated as a standard object of volume. That being said, if your source volume is LunA with 10GB capacity and you take a snapshot. The snapshot say LunASnap will be its own container with 10GB capacity. 

I assume your real concern here is physical capacity utilization of the array. If that is the case, there is no mechanism to stop snapshot creation due to capacity state. Again the snapshot does not consume physical capacity, it only uses shared memory for tracking the increased pointers to the deduplicated data.

The array does include default alerts for physical capacity utilization. 

Assuming you are on code greater than 4.0.x, alerts are raised at following points.

For XtremIO 4.0.x and greater, the following indicate the cluster has reached a capacity utilization of 85% or greater:

  • Cluster has exhausted 85% of its available physical capacity: Sys Ud Ssd Space Limited
  • Cluster has exhausted 90% of its available physical capacity: Sys Ud Ssd Space Very Limited
  • Cluster has exhausted 95% of its available physical capacity: Sys Ud Ssd Space No Free

These alerts are meant to trigger storage administrator actions. It is recommended to run regular storage reclamation activities from host side in order to restore space from deleted data from host perspective. This regular activity is preferred to a last minute scramble due to rapid capacity utilization. If the array hits 100% of physical capacity it will go read only. Any new writes will not be committed.

Note: space reclamation is covered in the XtremIO Host Configuration Guide 

https://support.emc.com/docu56210_XtremIO-Host-Configuration-Guide.pdf?language=en_US

Please note the alert information was pulled from the following knowledge base article. 

468164 : XtremIO: Managing Cluster Capacity in Response to Product Alerts https://support.emc.com/kb/468164

 

I hope the above provided the insights and information you need in regards to your query.

Best Regards

David Simpson

0 Kudos
Reply