Unsolved
This post is more than 5 years old
10 Posts
0
19462
Powervault MD raid rebuildng
Hello,
I have a Powervault MD which is connected to a PowerEdge 2950. the Powervault MD is in a raid 5 configuration with 1TB drives.
I noticed that one drive is in a 'drive failed' state and another is in a 'predicted failure reported by drive'.
I went to replace the 'drive failed' drive and the replaced drive went to solid green 'drive is online' state instead of going to the rebuild mode. In previous HDD replacements that I have done directly on other PowerEdge 2950's, the drive would go into rebuild mode right after inserting it. Is it not rebuilding because one of the drives is in 'predicted failure' mode? Or do I have to manually tell it to rebuild being that its on a attached Powervault MD?
Thank you for taking the time to read this.
DELL-Daniel Ca
243 Posts
0
April 16th, 2015 08:00
Hello, Henrymayes .
In this case, you should open 'Open Server Management Administrator' (OMSA), and set the drive you just replaced to be a 'hot spare'. It *should* start rebuilding soon after.
As for the predictive failure, it just means that drive has reached a number errors that would determine it's going to fail soon. The controller figures if the drive continues to encounter errors at the rate at which it's currently encountering them, it will fail soon. I would recommend replacing that drive soon as well.
I hope this helps. Let me know if you have any questions.
Henrymayes
10 Posts
0
April 16th, 2015 13:00
Hi Daniel,
I really appreciate your advice. I recently inherited the duties of these servers and there is very little documentation on them or even what the data each one is hosting. They have been running for about 4 years and I just installed OMSA for the first time on them.
Upon installing the new HDD, OMSA reports the drive as failed still, even though on the device itself it reports as online (solid green). Even hours later after the install, it still shows as failed in OMSA.
Previous raid arrays I have replaced drives on have started reuilding right away. But those were directly installed onto a server whereas these are in a attached Powervault MD1000, if that makes a difference.
I went to try to manually rebuild the array by clicking on "Storage" and the drop down "Available Tasks" window is not there only the "Select Reports" drop down. I do not see an option to reassign the drive as a hot spare either (although as far as OMSA is concerned, the drive is failed).
Per the list of the disks in the effected virtual disk array, it seems to not be using disks 5, 10 and 11. But, they don't seem to be set as hot swaps.
Some additional information on the array:
Storage page:
Physical disks list:
(Disk 9 blinks green/amber "Predicted failure reported by drive". Drive 11 shows steady green. Drive 12 was giving blinking amber "drive failed" but was the one that was replaced and shows steady green on the physical status indicator of the MD1000, but still shows as failed in OMSA)
Virtual Disk list:
(If needed the cut off information above is "Stripe Element Size" = 64 KB and "Disk Cache Policy" = Enabled)
Specific Virtual Disk:
Henrymayes
10 Posts
0
April 17th, 2015 14:00
Update:
After leaving the newly replaced disk in slot 1:2:12 overnight, now when I look at the OMSA interface on the specific virtual disk it now has removed 1:2:12 from that virtual disk. I thought it was weird that the virtual disk was initially missing 3 random disks when I first looked at it.
I am beginning to think that prior to my arriving, somebody replaced disks 1:2:5, 1:2:10 and 1:2:11 and for one reason or another the array was never rebuilt on those replacements (maybe 2 failed disks were replaced at once?). Somebody physically replaced the disk and it lit solid green like mine did and they just walked away without verifying the array was going to rebuild. Then the disks just got removed from the VD list and after the 2 disks were lost this way, the array failed.
Does that seem to be the case here?