Unsolved
This post is more than 5 years old
2 Intern
•
137 Posts
2
719
Changes in FAST VP Tiering behaviour in latest OE
Hi everyone,
one of our customers has noticed that the amount of data to be relocated by FAST VP has increased a lot since the latest Block OE code was applied to their VNX5600.
Has anyone else experienced this? Here's a sample screenshot from M&R.
The rise happened just after the Code was updated to 05.33.009.5.155 (from 5.33.008.5.119). The customer has observed similar behaviour on 2 different VNX systems, at different locations, with different workloads. Both arrays were updated at different times, but the change in behaviour was only after the upgrade.
I have checked the release notes and asked support, but could find no details about this.
Thanks and regards,
Stephen
turri01
1 Message
0
October 14th, 2016 17:00
Hi
Do you have 3 tiers in your pool and could you tell if it migrated up or down in the tiering.
umichklewis
1.2K Posts
0
October 17th, 2016 07:00
That seems a little too suspicious to happen on two different VNX after the code upgrade. Can you engage your local EMC SPEED Guru and make some inquiries? They might have an idea about an implementation change or other best practice that might be necessary on this code release.
Let us know if that helps!
Karl
SteveK821
2 Intern
2 Intern
•
137 Posts
0
October 17th, 2016 08:00
Thanks for your replies. I have already opened an SR and it has been escalated to engineering.
The pools have 3 tiers, and most of the movement is between the performance and capacity tiers, in both directions.
What I have also noticed is that the pool usage on one system is also increasing (free space going down, even though the amount of data had not changed). Customer said they cannot pinpoint any attached server with increased usage. And this also started around the time of the code update. Very strange.
I have a funny feeling that Deduplication is the culprit. One system has 2 pools, one with some dedup LUNs, the other without (as far as I can remember). The pool with Dedup LUNs is the one with the increased data movement.
I will keep you posted how support gets on.