vSphere reports HDD Predictive Failure

[mnickels]

Hello,

I have an HP DL380 G5 Server with two arrays, drives 7 & 8 are for the OS (RAID 1), 2-6 are data (RAID 5), and 1 is a hot spare for 2-6. I have vSphere 5.0 installed on this server, as well as the HP bundle with CIM providers.  All firmware updates are applied.

Drive 5 is giving me a warning status of predictive failure on my vSphere client.  The HDD is still flashing green like normal on the server - there are no amber/blue lights.

I read that it would may be better to wait until the HDD has officially failed before replacing it.  This way the disk can be spun down, and my hot spare would activate.

I've also read that it can take a long time for a drive to fail, and I obviously don't want two to fail at the same time.

What is the best course of action?

1) Wait until the drive has failed

2) Pull and replace the drive now

3) Is there a way to manually take the drive offline without restarting the server? This way I can replace the drive while it's spun down?

I do have DR backups, but obviously I would prefer to not have to use them.

Thanks,

- Mike

[SebastianGrugel]

I always for sure replace disks if have that warning...

"Is there a way to manually take the drive offline without restarting the server? This way I can replace the drive while it's spun down?"

Is it disk "hot spare" ?  if hot spare ( hot swap ) can be replace during working server without restart i think.

But maybe somebody more experience say something more :) i don't have to many experience.

Sebastian

[mnickels]

Thanks for the reply,

The failing drive is not the hot-spare.  I called HP support and he told me I could take the old drive out and put the new one in without issue.

I'm doing a few backups tonight, and I plan to replace the disk early tomorrow morning.  I'll post back with the results!

[SebastianGrugel]

OK. I will waiting for information from you :)

[mnickels]

Everything went great, details below!

The drive was still in predicitve failure status when I came in. I shut down as many non crucial VMs as possible to minimize disk reads/writes during the rebuilding process.  I then pulled the HDD, making sure I pulled the correct one from the correct bay.

On my VMWare Infrastucture client I saw that it was using my hot-spare to rebuild.  I immediately replaced the failed disk with a brand new drive and after a few minutes it started using that one to rebuild instead of the hot-spare.

I checked on it an hour later and everything is green!