Has anyone experienced the behavior where the backend ( VMAX 20K) gets very busy for up to 24+ hours after an upgrade to 5876. We recently upgraded an array that averaged 10% util and spiked to 60-70% for 24+ hours.
Don't worry about the DA utilization. If it was 10% before, it is still really ~10% now. The way FAST VP background tasks are scheduled is planned to change with the next release.
Are you talking about DA utilization? There was a change in the way the DA scheduler mananged FAST VP tasks, allowing more time spent on that when the DA was idle.
YEs,.,,I am talking DA util. We have many more arrays to upgrade and I cant afford that spike in the backend . My customers were bitten with the 5875. 249 issue where the data all nosed dive into SATA.
We had 24+ hours of higher than normal activity. We are tracking it down to see what caused it. This spike will impact my customers on other arrays. EMC has lost alot of credability with the 249 upgrade.
We are also running 5875.249.188 and had no issues yet but FAST VP is not heavily used. What temporary fixed were applied? We have epack 0061 which includes a couple of hotfixes. Can you refer to a Primus?
Quincy561
1281 Posts
1167
0
Posted September 24th, 2012 14:00
red1918sox2
Don't worry about the DA utilization. If it was 10% before, it is still really ~10% now. The way FAST VP background tasks are scheduled is planned to change with the next release.