Replacing failed disk - VNX 5200 Refusing to use it

Question

I had an apparent failure of a 600 GB SAS disk, P/N 118032757-A02 (Seagate Savvio 10K).  My vendor shipped me a 600 GB 10K SAS disk replacement, but it's a Hitachi, P/N 118032767-A04. When I plug the new disk into the same slot as the old disk, the light blinks blue momentarily, then the orange light illuminates, and the VNX sees the slot as empty (and if I re-run the Replace failed disk wizard, it shows as an eligible disk to be replaced).  I bought 3 disks from the vendor, and 2 of them did the same thing - the 3rd didn't light up at all, so I'm assuming it's a DOA disk. Am I missing a step, or do replacement disks have to be completely identical, or is it possible I somehow got 3 bad drives, or is there something else that could be going on?  I've never had a failed disk in my VNX before, so I'm really not sure what the process should be.  Could there be some sort of firmware issue on the disk (as in, they've got newer firmware than my array), or something?

kelleg · Accepted Answer

Please see the following Guide to the version of Flare running on your array and which version supports which Disks. Some new disks will not work with older Flare.

https://support.emc.com/docu42949_All_VNX_CLARiiON_Celerra_Storage_Systems_Drive_and_FLARE_OE_Matrices.pdf?language=en_US

glen

jwadmin · Answer

Okay, so it turned out that the disk I received was almost, but not quite, compatible. I've received an exact match for the failed disk, and so the VNX seems satisfied with it.

But: It's showing as "Unbound" in the Disk list under System -> Hardware, and so I'm not at all convinced that the system is actually willing to use it for anything.

I understand that the VNX2 permanently assigns hot-spares as the replacement when a disk fails, so I don't expect this drive to resume the exact role that its predecessor had, but I do want to make sure it's available as a hot-spare for the next failure.

When I go to the storage pool and hit "Expand" I just see a blank window (like, literally just the titlebar and the window frame...nothing ever draws inside the window). So to me, this seems problematic, like something is timing out and failing to return the necessary information to the Unisphere application.

I have no idea where to go next.

kelleg · Answer

Yes - the new disk is now considered for use as a hot spare and nothing more needs to be done on your part - job completed. glen

kelleg · Answer

On the VNX2 series, when a disk fails, one of the unused disks on the array (the ones that you do not use that are used as hot spares) is used to replace that failed disk in the Raid Group or Pool. Once the data from the failed disk is copied to the hot spare, the failed disk can be replaced. The new replacement disk now becomes a new potential hot spare. It is unbound because it's not currently being used.

Read the White Paper below - page 62 Mulitcore Raid - Permanent Sparing section for an complete explanation.

https://support.emc.com/docu48708_Introduction_to_the_VNX2_Series_-_A_Detailed_Review.pdf?language=en_US

glen

jwadmin · Answer

So then, I don't need to explicitly add the new disk to the RAID group or the pool?  It can (and should) just hang out in an unused, unbound state?  If so, then I guess I'll consider the job done.

imadblg · Answer

hello, i have one question about replacing faulthy disk on vnx,so can i replace the disk hot swap without i run Replace failed disk wizard on unisphere?

DELL-Josh Cr · Answer

Hi,

Thanks for your question about replacing a disk. It should make it a hotspare automatically and allow it to rebuild. If it isn’t working make sure the disk is for VNX, disks without the proper firmware won’t work. Also make sure that the drive is large enough and not a few GB smaller.

Let us know if you have any additional questions.

VNX

Was this post helpful?