Help with PERC 3/DC Mirror Array

Assuming these are brand new, not refurbs/recerts, and you have the latest firmware install on the card, are the firmware revisions on the drives exactly the same ? Has the disk initialization finished before you reboot ( been ages since using a 3 DC)?

The drives where previously used, not brand new, but they are detected by the PERC ok so I am currently assuming their worthyness.

Would it be worth attaching them to the server's SCSI controller first to see if I can able to install ESXi on them individually to verify that they are working correctly? Would it also be worth low-level formatting them via the server's onboard SCSI controller to verify that there is no configuration being left on the disk by the PERC formatting process?

The two drives are the same make/model/capacity and they both have the same firmware revision running on them. The PERC 3/DC is running firmware 199A which I updated from the Dell site yesterday before I started to use the card as it had a really old version on it.

The initialization had completed and I had the configuration changed and reboot required prompt on the screen before I restarted the server.

Thanks
Richard

I've verified the problem is not the controller by creating a Mirror using two 36GB Fujistu MAP drives I had as the array build and ran correctly.

I've moved the two 73GB drives to the server's on-board SCSI controller and am low-level formatting them using the SCSI-Select utility to ensure that they are blank with no legacy configuration and will then try them on the controller again.

With regard to media errors, the PERC showed no media errors for either of the 73GB disks whilst they where attached to the controller, so at this point I can only assume there is some configuration on the disks which the PERC clearing process did not remove - Perhaps a configuration from a non-PERC RAID device used by the previous owner??

Ok - Drives are still not working and I'm starting to get annoyed with them now.

I've just taken the PERC firmware from 199A back down to 1.72 and then back up through all revisions in between to test the drives at each level and I get the same error all of the way through.

Is it possible for a disk to be incompatible with the PERC? I'm wondering why the drives show no media errors and appear generally fine, but yet they will not work in RAID even though nothing is wrong with either the on-disk configuration or the NVRAM configuration - They both match and are valid?

Is it possible my stripe size or read-through settings are causing the problem? I am using the default settings of 64KB stripe and Adaptive read-through.

I would really appreachiate (spelling?) some help with this.
Thanks

Sorry for the delayed answer, I answered this yesterday but the post disappeared.

"detected by the PERC ok so I am currently assuming their worthiness."

Drop the assumption. In my earlier years working with raid arrays, a drive would fail and I would place them on a standard SCSI adapter and continuously test for two weeks. Very often the drive would pass with flying colors, no errors at all, with multiple test suites. Place the drive back in an array, it would fail. Again place the drive back on a SCSI adapter, load up an OS, and it would run for years with no issues. After seeing this a number of times, I do not test anymore, I replace; in your case purchase one drive, replace either ,give it a try, if no go, put back the replaced drive, and swap out the other.

" Is it possible my stripe size or read-through settings are causing the problem?"  Not possible

"Is it possible for a disk to be incompatible with the PERC?" Possible, but unlikely, even if they show as no errors, there could be other reasons, possible one the chips on one of the drive controllers is going off spec. You might google the drive model and see if there are any hits as to compatibility. 

"I would really appreachiate (spelling?) some help with this." spelled appreciate  :emotion-2: , if it were not for spell check, I would be lost

So I took your advice and just picked up some replacement drives this week which arrived today.

To my dismay, these drives also gave the same mismatch error message. I cleared the drives and cleared the controller configuration and rebuilt the array and the logical drive, but upon the restart, I got an error message which looked slightly different, but along the same lines.

I'm running the Consistency Check at the moment to see if that pops anything, but as with last time - no media or other errors reported from the drives.

FYI: The drives I've bought this time are Fujistu MAU drives which I decided to get based on the fact that my two Fujistu MAP drives worked in a mirror configuration on this controller without any error, however it seems that MAP and MAU are not built equal.

I did a search online earlier also just to make sure I wasn't being an idiot, because the new drives I am buying for the mirror are U-320 disks, where as the PERC 3/DC is only U-160 as are the MAP drives which worked fine. I checked as I said that I wasn't being a fool and everywhere I looked told me that U-320 is backward compatible with U-160.

Does the PERC 3/DC have really bad compatibility problems with drives or am I just shite out of luck on this one picking up four bad drives in a row? I'm starting to regret deciding to run ESXi on this machine - ESXi is the only reason I upgraded from a PERC 2/SC to a PERC 3/DC because the 2/SC isn't supported nor is it detected by the ESXi installer.

Odds of 4 drives being defunct, extremely unlikely. Personally I have never tried u320 in a 160 system, but being the u160 drives are difficult to obtain, I  would have tried the the u320 drives... at the same time I would be worried even though they are "backward compatible". When I was using these adapters I was able to get the exact same drives, which were all Seagates; doubt  Fujitsu should make a difference, beside you were able to get one set working.