3par 8400 SSD raw space usage above 95 % - How to reclaim space ?

[FS-EXPLOITATION]

Hi,

We've been getting an alert lastly about the threshold reached by our SSD virtual volume.

The message code is 0x027001c and it says SSD raw space allocation has reached 95 % and soon all write will fail.

Does someone know what actions can be done to alleviate the problem without buying new disks?

Also, we've been wondering how a virtual  volume allocate space when there is a mix of differents disks in it like SSD and FC.

Thanks for your help,

OHC

[SUBHAJIT KHANBARMAN_1]

@FS-EXPLOITATION 

Check your CPGs and verify if any space can be reclaimed by running compactcpg in dryrun mode.

compactcpg -dr <CPG_name>

If you find GBs of space to reclaim then you can execute the same command without "-dr" option but if you find TB of space need to reclaim then better to do the same in Out of office hours.

Do you use Adaptive Optimization feature?

There are many ways to do space reclaim. If you have deleted snapshot or VVs directly from 3PAR then space reclaim will be faster.

If you have deleted data at the host filesystem level then space reclaim need to initiate from Host (depends on Host OS and it can be automatic as well). This method is slow.

Hope this helps!
Regards
Subhajit

I am an HPE employee

If you feel this was helpful please click the KUDOS! thumb below!

**********************************************************************

[Anjum_IK]

@FS-EXPLOITATION 

Good day!

"The system has less than 5% of the SSD-free capacity remaining. The available space needed for auto-growth of logical capacity by CPGs is at risk. It is critical to add more physical capacity as soon as possible. If all available space becomes full, new writes on all virtual volumes will fail and data will become unavailable. (which you are aware of already)

The system has used over 95% of the raw SSD physical space. The number one cause of outages on arrays.

NOTE: A CPG with customized LD characteristics can be limited to using a specified subset of total available space. That subset can reach these thresholds of percent full even when other physical space is still available.

NOTE: Sometimes a single allocation hits both the Allocation Warning and the Allocation Limit at the same time. When that happens, only the more severe alert—the Allocation Limit—is posted. For example, if the Allocation Warning is set at 80% used and the Allocation Limit is set at 90% used, then if you create a single VV that takes you from <50% used to >95% used, the Allocation Warning alert is bypassed and only the Allocation Limit alert is generated.

Review the suggestions below for ways to make more space available:
Evaluate reclaiming space from CPGs.
Evaluate reclaiming space within TPVVs.
Check for LDs that are not mapped to a VV.
Identify volumes that are exported to a host that may no longer be necessary.
Evaluate snapshots as they can use more space over time.

Run Compactcpg and run freespace commands to reclaim the space on the array.

[Sheldon Smith · Accepted solution]

we've been wondering how a virtual volume allocate space when there is a mix of differents disks in it like SSD and FC.

It sounds like you are using Adaptive Optimization.
If you are, do you also have virtual volumes using a non-AOmanaged SSD CPG? That is, a volume that is SSD, thin, and not using an AO CPG?

If all your volumes are under AO control, you don't care.
If AO is unable to move regions from FC to SSD, it won't.
And just set the RawSpaceAlertSSD to 10 to silence the alert

If you do have one or more volumes that are all SSD, you may have to adjust the AOcfg to limit how much SSD space is being used by AOmanaged volumes.

This has been discussed recently. Search the forums.