Dell compellent SC 8000 performance optimization

Issue :

Performance slow - due to DC storage Tier 2 Raid 10 Full, sync replication data write in lower performance disk caused overall system performance slow.

Incident :

15^th Oct 2018

-       We have fixed support assist and now Dell has received performance data from DC [ SSN 40249/50 ] and DRC1 [SSN 40251/52]

-       After review performance data and found out DC storage -Tier 2 Raid 10 full which is caused the performance impact

-       Change some of the replication volume in DC storage from original Tier 2/Tier 3 storage profile to Recommended storage profile

-       Apply snapshot to those replication volume

-       Manual kick start data progression to move data out from Tier 2 Raid 10

16^th Oct 2018

-       Explained about how sync replication works

-       Change all sync replication volume [ DC ] to use same storage profile as DRC1

-       Apply snapshot 25 hours expiration for all sync replication volume in DC

-       We have check so far during peak time – 11am to 2pm no high latency like previous. Please continue to monitor

Pending action items :

Secure console access.

Recommendation :

Please continue monitor if Tier 1 full then please change storage profile to T2/T3.

As discussed, synchronous replication is very sensitive and if any of the source or destination use different type of disk then performance may not be optimal

Here is the summary for today.

You have a lot of view volume on remote system using T1 T2 T3

Those one had no replay profile set on the volume

So all data where stuck on the T2 raid 10 and this one was full and unable to move down data

So destination volume was unable to write on T2 raid 10 (lack of space) and so all write come into T3 .

So it generated a big latency on synchronous replication.

Action done :

Assign a replay profile to those volume  (Dataguard)

Create manual replay to free up space on the T2 raid 10

Launch a manual DP

Change some storage profile to recommended on the destination array to remove some load on the T2.

Next actions :

Monitor the system T1 and T2 free space

If the used space on T1 reach 80 %, change back some volume to the T2 T3 storage profile

If the free space on the T2 raid 10 reach 50% change back some volume to the T2 T3 storage profile

At the end you will be able  to change back the configuration after the space on the T2 raid 10 will stabilize