RAID management - Predictive disk SMART failure

activateahsd
activateahsd used Ask the Experts™
on
Hi,
I have an HP ML370 with a P400 raid controller installed. It has 6 disks in a raid 5 config. 5 active disks and a 6th spare. The ACU is reporting a SMART predictive failure event for one of the 5 active disks (the server is working normally currently).
My questions are, do i assign the spare disk in the ACU to the array? what will happen to the array when i do this? will the spare be automatically included in the array once assigned and how will i know when it is safe to remove the disk that has the predicted failure?
Many thanks
Andy
Comment
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
Top Expert 2014
Commented:
No, the hot spare will not be used for predictive failure, just real failure.

It's always safe to remove the disk with predictive failure and replace with a new one. If you remove it hot it will start to rebuild onto the hot spare assuming that's assigned to this array, then you fit the replacement and it rebuilds back onto the replacement.

Alternative to removing hot you can shutdown, remove, power on, replace - do not fit the replacement with the server turned off.

Another alternative if you have the replacement disk to hand is to temporarily unassign the hot spare from the array, then pop the failing one out and replace. That stops it wasting time rebuilding onto the spare then rebuilding onto the replacement again. Don't forget to assign the hot spare again for those long weekends where there's nobody around to replace failed disks.

Also check the firmware of the failing disk, it may not be failing at all since HP re-tweak the SMART predictive failures on some due to false predictions. I can check if you post an Array Diagnostic Report (same place as ACU in menus) as an attachment. Don't paste as text in thread, it's huge.

Author

Commented:
hi - thanks so much for looking at this...
Is the spare in the array and ready to kick in automatically? It is a listed disk under 'SAS Array A with spare' in the ACU
report.zip
Top Expert 2014

Commented:
Yup, the spare is assigned to Array A so will kick in as soon as the one with predictive failure is removed. It doesn't say why it's predicting failure but I checked the firmware revision and although there is later firmware it's listed as optional rather than required and doesn't address any SMART problem.
Bootstrap 4: Exploring New Features

Learn how to use and navigate the new features included in Bootstrap 4, the most popular HTML, CSS, and JavaScript framework for developing responsive, mobile-first websites.

Author

Commented:
Hi - thats great, thanks very much for your help.
The rebuild priority is set to medium, is it best to remove the failing drive at the end of the day due to performance hits? Approx how long does it take for the spare to be fully built (array is 280GB)
Andy
Syed_M_UsmanSystem Administrator
Top Expert 2011

Commented:
you have two options;
1)      since you have identified the faulty Disk, simply remove faulty Disk plug new disk, your raid card will automatically generate parity (dont use spare in this case).
2)      Go to HP Tools-¿Smart array configuration utility>Select your Raid card--->from menu you will be able to add/remove Disk is existing logical drive

if you are not expert i suggest you dont use Option 2.
Syed_M_UsmanSystem Administrator
Top Expert 2011

Commented:
sorry for late reply, i fully agree with andy.
i opened this question when author asked the qustion but got stuck with someone, and when i posted my reply i found so many experts reply.
Top Expert 2014

Commented:
Could take about a day to rebuild, 4GB/hour was the old rule of thumb but it's a bit faster than thet nowadays. Rebuild priority doesn't make much difference if the server is idle anyway, if it's set to high then rebuilding taks preference so user performance decreases. If you have the replacement I'd be more tempted to remove the spare to avoid it doing the operation twice.

Author

Commented:
superb answer - thanks so much!

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial