vpnsol123
asked on
Predictive Drive Failure replaced - performance dropped
Hi All -
We have a Dell R720 with 8 300GB 15k SCSI drives in a RAID 10. Recently one of the drives reported a predicted failure. Purchased a new 300GB 15k drive and pulled the predicted one and replaced with the new. The RAID started the rebuild and completed successfully. From the Dell Open Manage program - you wouldn't know anything was wrong. RAID is healthy and disk shows happy. What we noticed was that after that - performance dropped. (Its an app server with a few IIS apps and runs SQL)
End users were stating they were getting timeout messages and such and I could tell in the OS it felt like it was dragging. I took a look at the history of the SQL backups prior to the drive placement and post placement and it was indeed 5-10 times longer to do simple transaction log backups so something was up.
I popped the new drive out of the chassis (which of course degraded the RAID) and found the speed came right back and its quick again. I've not had any issues replacing bad drives before where the performance changed post rebuild with the same size and speed drives.
I'm guessing upgrading the PERC H310 Controller firmware may help - it does show out of date (and I'm pretty sure it was previously) and if that doesn't give any relief - purchase a second new drive? Sound about right?
We have a Dell R720 with 8 300GB 15k SCSI drives in a RAID 10. Recently one of the drives reported a predicted failure. Purchased a new 300GB 15k drive and pulled the predicted one and replaced with the new. The RAID started the rebuild and completed successfully. From the Dell Open Manage program - you wouldn't know anything was wrong. RAID is healthy and disk shows happy. What we noticed was that after that - performance dropped. (Its an app server with a few IIS apps and runs SQL)
End users were stating they were getting timeout messages and such and I could tell in the OS it felt like it was dragging. I took a look at the history of the SQL backups prior to the drive placement and post placement and it was indeed 5-10 times longer to do simple transaction log backups so something was up.
I popped the new drive out of the chassis (which of course degraded the RAID) and found the speed came right back and its quick again. I've not had any issues replacing bad drives before where the performance changed post rebuild with the same size and speed drives.
I'm guessing upgrading the PERC H310 Controller firmware may help - it does show out of date (and I'm pretty sure it was previously) and if that doesn't give any relief - purchase a second new drive? Sound about right?
Seth Simmons
you need to take the disk offline first before replacing when it is only a predictive failure
ASKER
Hrm. So now that step wasn't done - any suggestions on how to proceed now that the new disk isn't a predicted failure unless I need to put the old disk back in - let it rebuild - probably will say predicted failure - take it offline - then replace the drive with the new again?
yeah i think that's what i had to do at my last place
HP controllers you don't need to do that but PERC needs to have the disk offline first
HP controllers you don't need to do that but PERC needs to have the disk offline first
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
I ordered a dell certified drive and put it in - performance was restored