2 Posts
0
701
November 7th, 2022 00:00
Dell PowerEdge 2900 Failed Physical Drive (Raid 5)
I have a Dell PowerEdge 2900 server running Windows Server 2003. There are 3 SAS drives in the system configured to operate in Raid 5. One of the disks is failed and another one is blinking green and amber which means predicted failure.
I am concerned about how to proceed for multiple reasons
The server is running a specific software that cannot be installed anymore and I haven't found a way to make working backups either. The system itself is in a bad state, generating tons of erros because everything else than the said software has been moved to new devices. This software is the only reason the server is still operating.
I figured my best bet to keep the server working as long as possible would be to replace the failed disk with a working one - but I am not sure if it would work. I am unable to get the same type of SAS disk which are currently used. I am able to get a SAS disk bigger in size and slower in speed, only an used one though which might or might not have unknown data on it. I do not know if the "new" used disk having data on it would impact the rebuild if I hot swap the failed drive while the system is running. I also do not know how much the rebuild relies on the server operating system - because if it is not completely done by the PERC and needs a fully working system i wouldn't be surprised if the rebuild fails. There are files that are either corrupted or being used all the time - I can't even copy them to an external drive.
So I am asking for advice on what should I do. Maybe not touching it and hoping it will keep working as it is would be my best option? If I am supposed to replace the failed drive what is the correct way? It it best to hot swap while the system is running? Should I boot into the PERC Configuration Utility and replace the disk from there?
Also this might be a question for a different category but if anyone can suggest a reliable way of creating a backup I would appreciate the information. The built-in backup and restore wizard worked for creating a backup file but I haven't been able to restore it on a virtual machine.
theflash1932
11 Legend
•
16.3K Posts
1
November 7th, 2022 09:00
Replace the failed disk. Any 3Gbps / SAS-2 HDD will be just fine - it can be larger or faster / slower. Insert it "hot" and it should start to rebuild. If not, then it likely has flagged the drive as "foreign", and you will need to clear the foreign configuration first then assign it as a hot-spare to start the rebuild.
Once the drive has rebuilt, you'd be wise to replace the predfail disk as well.
After repairing the failed drive(s), come up with a way to back it up ... you can use the built-in NT Backup. Another option is to convert it to a virtual machine, so it is running on solid hardware, then you can keep it backed up with modern software.
THEN you need to come up with a plan and a timeline to replace it with new software. From the sounds of it, this will NOT be cheap, but it is a necessary step in preserving the continuity of the business. This is a cost of doing business, and any responsible business owner should be planning how to make these types of upgrades and expenses possible.
DiegoLopez
6 Operator
•
2.7K Posts
1
November 7th, 2022 08:00
Hello @gjkb886,
It's difficult for me to be able to offer you some advice regarding this issue. But, what I am sure is that you should not "not touching it and hoping it will keep working as it is would". A RAID 5 with 2 failed drives will not be accessible and all data will be lost. And if one disk has alread failed and the other one is in pre fail. This situation will happen sooner or later.
Your best option is to replace first the failed drive and second the pred fail drive. The rebuild process in the first new disk could be ok even when a disk in pred fail.
Regards.
gjkb886
2 Posts
0
November 15th, 2022 01:00
Thank you for the advices. Thanfully replacing the failed drive worked. The light is now steady green so the rebuild should be completed.
For the predicted failure drive, I should just hot swap it the same way?
The built-in NT Backup worked for creating a backup file but I haven't been able to restore it to a virtual machine, it simply stops doing it's thing at some database files. I haven't found any working way so far to convert it to a virtual machine.
DiegoLopez
6 Operator
•
2.7K Posts
1
November 15th, 2022 05:00
Hello @gjkb886,
Basically, a predicted failure is caused by the drive having exceeded a threshold on bad blocks. This means the disk will soon fail. To avoid this from happening on a random, not controlled moment. You should replace the drive as soon as possible.
In order to replace it without transferring those bad blocks back to the disk you will need to force the predicted failure drive offline, then remove it and replace with new drive. You can force offline or power down the server and remove the drive with server off (Do not add replacement while server off or that can harm data).
Regards.