Promise Technology

Promise Technology :: RE: best practice for replacing a failed disk

abstractrude — Thu, 23 May 2013 05:50:20 GMT

Author: abstractrude
Posted: Thu May 23, 2013 12:50 am (GMT -5)

that also works. you just have to do two data moves that way. If you have a backup and wanna get rid of drives quickly i prefer just doing as fast as possible but PDM will also help solve this problem. PDM is definitely safer if you don't have a backup.

I havent relied on PDM much, but the idea is PDM removes rebuild times by migrating in prefailures. in this case you just need to replace suspect disks now.

Promise Technology :: RE: best practice for replacing a failed disk

arls1481 — Wed, 22 May 2013 18:34:26 GMT

Author: arls1481
Subject: RTFM
Posted: Wed May 22, 2013 1:34 pm (GMT -5)

NM!

Somebody should have just told me to....
Using PDM now to migrate, once thats done you swap out the suspect disk with a new one and flip back your revert spare after you initialize the new disk to your LD.

Yay for RTFM!
Yay!

Promise Technology :: RE: best practice for replacing a failed disk

arls1481 — Wed, 22 May 2013 11:11:38 GMT

Author: arls1481
Posted: Wed May 22, 2013 6:11 am (GMT -5)

Thanks!

So it's marginally safe to leave a shared volume running whilst I do this?

Promise Technology :: RE: replacing 750gb drives with 2TB drives in Promise Vtrak

xsanguy — Tue, 21 May 2013 23:41:48 GMT

Author: xsanguy
Posted: Tue May 21, 2013 6:41 pm (GMT -5)

Strangely, the absolute best 2TB drive may be a 3TB drive. Seagate 3TB barracudas seem to be quite reliable, and very very fast. Only 2TB will be recognized, but it will be the fastest part of the drive, out-performing any other 2TB drive made.

Promise Technology :: RE: best practice for replacing a failed disk

abstractrude — Tue, 21 May 2013 18:33:23 GMT

Author: abstractrude
Posted: Tue May 21, 2013 1:33 pm (GMT -5)

get rid of your hot spares. pull bad drive and replace wiht good drive. when done enable hot spares

Promise Technology :: best practice for replacing a failed disk

arls1481 — Tue, 21 May 2013 17:11:38 GMT

Author: arls1481
Subject: best practice for replacing a failed disk
Posted: Tue May 21, 2013 12:11 pm (GMT -5)

I've got a handful of suspect drives in my vtraks and I want to replace them with new drives but I want to know the collective thoughts on a best practice to do so. If it's of any consequence, the failing drives are all OEM 750GB SATA in vtrak e610f e & j class drawers and are at about 4 years of life-cycle so its time...

So here's my thoughts, please correct them as you all see fit:
1) mark a defective drive as offline
2) pull dead drive and replace with a new drive
3) transition hot spare to replacement drive???

Do I need to stop the volume before I do any of this?

Promise Technology :: RE: replacing 750gb drives with 2TB drives in Promise Vtrak

maxscience — Tue, 21 May 2013 12:09:47 GMT

Author: maxscience
Posted: Tue May 21, 2013 7:09 am (GMT -5)

johnny are you sure? The partitions of my promise E-610 are GUID not MBR.... When those units came out there were already Intel Macs and Xserves so the GUID partition table was the preferred choice.

What do you guys recommend as the best 2TB SATA drive for the Apple version of the E-610 Promise unit?

Promise Technology :: RE: E-Class Drawer Status Check LED is Amber

maxscience — Sun, 19 May 2013 18:28:27 GMT

Author: maxscience
Posted: Sun May 19, 2013 1:28 pm (GMT -5)

I had this problem as well and struggled several hours before discovering the solution (without calling support, I'm abroad). Basically this has to do with SATA drives having only one port, so the MUX on each slot (that card that connects to the drive) is doing the work of multiplexing between the two controllers.
When this type of drive failure occurs, the controllers go nuts in trying to connect to the drive so one of them goes into maintenance mode.
Promise even has a page on this. They make it sound as it is an unavoidable architectural fact but the result is that you completely lose the so hyped redundancy of the system. Ironically all controllers work fine but you end up with a downtime because you have two redundant controllers instead of one.

Short answer: a restart won't help, you have to manually replace the drive that is causing the issue.
You won't have this issue with SAS drives as they don't need the MUX in each slot.

Promise Technology :: RE: replacing 750gb drives with 2TB drives in Promise Vtrak

Johnny_0_o — Sun, 12 May 2013 14:18:32 GMT

Author: Johnny_0_o
Posted: Sun May 12, 2013 9:18 am (GMT -5)

Wouldn't that be the GPT partition table instead of the MBR that causes this issue? It is the case with windows osx and linux so it is probably the same thing with the controler that was not originally made to support anything but the traditional master boot record. I am wondering if they could come up with a firmware fix of some sort.

Promise Technology :: RE: replacing 750gb drives with 2TB drives in Promise Vtrak

xsanguy — Sun, 12 May 2013 01:40:43 GMT

Author: xsanguy
Posted: Sat May 11, 2013 8:40 pm (GMT -5)

Unfortunately just confirmed that SATA 3TB (Seagate) and SATA 4TB (hitachi) don't work (well they work, but come up as 2TB.) in the E-class.

Supposedly both the Hitatchi HUS723030ALS640 3TB SAS and the Seagate Constellation ES2 3TB Seagate is certified. No idea why it's essentially identical brother (the 3TB Barracuda with SATA board) is not being recognized as a full 3TB.

There's a 2.2TB barrier with some SATA gear, but if the SAS drives can break it, shouldn't everything?

*bump* for other real world experiences.