Hello,
Have an old power edge 840 running a raid 1 with two 750gb drives, one has failed. Have a bit of confusion about part 1 that Daniel referenced:
According to DELL-Daniel My on another thread about this controller, he said:
You will want to boot into the SAS5 controller BIOS and OFFLINE/FORCE OFFLINE the drive you are replacing before you swap it out. If you do not force the drive offline then when the controller boots up with the blank drive it will likely wipe all of the data across the array. The problem is that the controller will not know the drive is replaced and it will try to make the two drives identical, wiping out the data.
Offline the drive in the controller BIOS or OpenManage Server Administrator/Storage Manager
Shut down the server and swap out the drive
Boot up the server and set the new drive as a hot spare through the controller BIOS or OMSA
The drive will start rebuilding
=====
I found a great step-by-step article but it conflicts slightly with what Daniel says so want to make sure I understand before attempting these steps:
sheldonsblog.com/index.php/rebuild-degraded-raid-mirror-on-dell-poweredge-840-with-sas5ir-controller/
1) -When- do I remove the missing drive (bad drive)? Do I boot into bios as Daniel said and force the "missing" (bad) drive to show as offline, shut down the system, then replace it with the new blank drive (is it possible to offline drives in the raid bios on this controller, if so, how)?
2) Or do I simply identify what drive is bad in the raid controller bios (labelled as "missing"), shut down the server, remove the bad drive, put in the new blank one, boot up & re-enter bios using ctrl+C, set the blank drive as a hot spare, and wait till it's 100% resynced as suggested in the sheldon guide?
I've not seen anything in the official Dell documentation about setting 1 of the 2 drives in the virtual disk (raid 1 array) as offline. Obviously on reboot, I don't want it to start syncing the new blank drive over the existing good drive, wiping out my data. Will have a backup beforehand. Doesn't the raid controller have info it references and the "degraded" drive have headers referencing the raid 1 array? Shouldn't it know not to sync the blank drive (without raid headers) to the drive with "raid headers"?
As such, can't I power down, remove the bad drive, put in the good one, boot into bios again, set the new blank as hot spare and wait till the sync is finished?
I know anything can happen and I have backups but trying to avoid wiping things out by accident.
Thank you.