Unsolved
This post is more than 5 years old
4 Posts
0
1775
HD Predicted Failure on Raid 5 Array - PERC S300 Controller - How to force offline for replacement?
We've got an older R310 server with a PERC S300 array controller in it. The array is configured in a RAID-5 setup. It's also got the hot-swap backplane/caddies. One of the drives is flashing green/amber which I know means that the drive is in a predicted failure state. (confirmed by viewing physical drive status in OpenManage Server Administrator)
I found the document for replacing a failed drive offline when using an S300 here: http://www.dell.com/support/article/us/en/19/sln290972/perc-s100-s110-s300-how-to-replace-a-failed-hard-drive?lang=en , but the doc says "This procedure applies only to replace a failed hard drive (FAILED state)." The drive hasn't failed yet, so I'm confused... How can I proactively replace this dying drive?
We've got a Dell certified drive on the way for replacement. How do I tell the server that the failing drive is offline? I don't have the "offline" option under Available Tasks when viewing the physical disks status in OSM. Do I just pull the drive while running and put the new one in its place? (Since the S300 supposedly supports hot-swap, is that enough?) Or, do I need to power off the server to add the replacement?
I'm thinking I need to do the following:
1. Power off the server.
2. Pull the failing drive.
3. Power on the server and enter CTRL-R before windows loads to enter the raid configuration. Hopefully, the server will notice that the drive failed?
4. Add the replacement drive and rescan disks so the controller can see the new drive.
5. Delete the new virtual drive that the controller will create from the new physical drive.
6. Add the new drive as a global hot spare.
7. Restart the server and let windows startup. At that point windows will start a rebuild using the new "hot spare".
8. Monitor rebuild status with OSM and when it finishes, remove the global hot spare designation on the replacement drive.
Thank you for your help!
jhowzr
4 Posts
0
February 12th, 2018 10:00
PM sent. Thanks!
DELL-Daniel My
Moderator
Moderator
•
6.2K Posts
0
February 12th, 2018 10:00
Not all of our controllers support the offline feature. Offlining an online drive before removal is best practice to avoid data loss/corruption, but if your controller does not support offlining a drive then the next best method would be to remove the drive when there are no writes taking place.
The safest way to remove an online drive without offlining would be to boot to the controller BIOS and hot-swap the drive. There should be no writes taking place while in the controller BIOS.
If you hot-swap the drive it should initiate a rebuild automatically. If it does not rebuild automatically then you may need to assign the drive as a hot spare. Also as you mentioned in your outline, the S300 may set the drive as non-RAID. If the controller does that you will need to delete the non-RAID disk that it creates in order to rebuild the replacement drive into the virtual disk.
https://www.dell.com/perc/
Thanks
DELL-Daniel My
Moderator
Moderator
•
6.2K Posts
0
February 12th, 2018 10:00
Hello
Please send a private message with your service tag to ensure we have all appropriate information on your system.
Thanks
jhowzr
4 Posts
0
February 12th, 2018 11:00
Basically, I need to remove the drive while the server is on (either in windows during low/no use, or in the controller BIOS) so that the controller makes note that one of the drives went offline?
Reading the PERC manual, I was lead to believe that the rebuild will only take place once Windows is running. So, if I do the swap while in the BIOS, will I get any indication that a rebuild will start once in windows, or do I just need to check on it once windows is back up?
I'm all for doing whatever is safest for my data! :-)
So,