ProLiant MicroServer Gen8 - Faulty Smart Array P212?

[STEPHEN PETTITT]

Hello,

Apologies if this is in the wrong place, but I did not see a section specifically for the MicroServers

I have an HP MicroServer Gen8 running Windows Server 2012 Essentials that, over the last few months, has started to exhibit an annoying issue.  It has an old Smart Array P212 installed and 4 x 4TB disks.  Periodically disks just "disappear".  They are not failing exactly, they just suddenly become invisible to the RAID controller, as though someone has just removed the disk.  Sometimes I can remove and re-seat the affected disk (offline) and then, when restarted, the controller rebuilds the array.  It can sometimes remain working like this for weeks.  On other occasions the disk disappears again during the rebuild process.

A couple of days ago, when using the array management tool, I received a message to say that, in effect, the controller itself was inaccessible.  This can be fixed with a cold restart.

With all of these issues, I have concluded, rightly or wrongly, that the controller itself is at fault.  I have therefore purchased a couple of replacement controllers, a like-for-like P212 and also a P222.

My question is, can I replace the existing P212 with the P222 now, even with one of the disks offline?  Or do I need the array to be completely rebuilt before it is safe to do this?  Should I use the newer P212 to rebuild it first?

Despite being an IT professional who has used HP servers for several decades, I have never actually had to replace a RAID controller before, so I am not aware of the exact procedure.

Any advice would be welcome.

Thanks in advance,

Stephen.

[krasinka1409]

Hi, 

the p212 is not supported in gen8 servers, so this cpuld be the cause of your issue. Additionally, non HPE brtanded HDDs may cause the same behaviour. 

It would be the best to rebuild the array completely first and than perform the replacement. Just power down the server, replace the p212 with the p222 and start. The RAID config is saved on the HDDs, so the new controller should recognize it, but in some cases, the RAID may need to be re-created. 

[STEPHEN PETTITT]

Thanks for the information.

The controller and its arrays used to be in an old G7 MicroServer.  When the PSU died on that I just decided to replace the whole server instead.  It never even occurred to me to check compatibility.  What's weird is that it was working fine for more than 6 months before it started misbehaving.

[krasinka1409]

Hello Stephen, 

yes, this does not surprize me. In most of the cases uncompatible hardware is funkctioning properly till any unexpected behaviour occurs. Better you install the p222. That would help at least to exclude the incompatibility as a possible cause. 

[STEPHEN PETTITT]

The saga continues.

I could not get the original P212 to be recognised at all, so the array was not able to rebuild.  I therefore thought that, as it would be a like-for-like swap, the least risk option would be to swap it out for the newly acquired, replacement P212.  I couldn't get the system to get past the POST screen with that.

I therefore bit the bullet and installed the P222.  It recognised the existing arrays (there is an external enclosure attached too) and so everything looked okay.  However, when the OS loaded and I was able to launch Smart Storage Administrator, it was still unable to "see" one of the disks.  So, I swapped it out for another, identical disk and on the next boot it was visible and the array started to rebuild.  However, I refreshed the screen after about 10 minutes and it had disappeared again!

Admittedly the disks are not HP-branded, because the extra expense of those doesn't make sense for a home server, but they are Western Digital Reds.  I would like to think that this is not a factor.

Does anyone have any other suggestions?

[krasinka1409 · Accepted solution]

Hello Stephen, 

actually the HDDs are a very imortant factor. Especially on gen8 servers. 

I have many customers experiencing the same behaviour using non hPE branded drives. 

[STEPHEN PETTITT]

I think I have found the problem.  It seems that the original disk and the spare that I had used are both suffering from similar faults.  I have tested them in an external eSATA enclosure attached to a PC and, although they both pass the SMART test, they are both exhibiting strange behaviour.  One of them spins up okay and then random stops spinning as though the power has been switched off.

I have temporarily used a Western Digital Blue in the array (as it was the only other spare that I had that was big enough) and everything rebuilt successfully.  Although it seems that I have unnecessarily upgraded the controller, the array did appear to rebuilt considerably quicker than it has in the past.  It's gone from about 24 hours to around 10, so it's not all bad news!