Jump to content
Corsair Community

V64s in RAID5 dropping like flies


timothym2

Recommended Posts

Built the RAID about 1.2 years ago. 4x NOVA V64 in RAID5 - initially on an Adaptec 3805 controller; later moved the drives to the motherboard ICH10 soft RAID5 controller.

 

Installed Windows 7, OS on the SSD RAID5, User, ProgramData, PageFile and Temp on the spinning disks, another 8TB RAID5.

 

A month or three back, disk 4 of the SSD dropped (newest FW/build of the four drives). RMA'd the drive, now have a refurb. While waiting for the RMA, I rebuilt the RAID5 using the 3 remaining disks. A few weeks ago I started having random crashes. No clue as to what was causing the locks - logs were unhelpful. Just a locked computer that needed a power cycle (reset didn't do it).

 

A few days ago, another full RAID5 drop, this time the same symptoms as a few months back: one of the disks would drop, a power cycle and the drive was back, only to fail again and no show in the raid manager.

 

So I've pulled the drives and tried reformatting them. One got to 96% then started acting up, but eventually finished formatting; the second dropped at about 40% during format (no longer showing in Disk Management); and the third seems to be having similar issues, and dropping around 96% (and disappearing from Disk Management).

 

1 of four was random - but now 3 out of 4 showing failures? Something seems odd about these drives. I really tried to ensure the drives were being used in a low write mode (only as storage, even moving the log files off the drives)... yet they seem to be failing in about 12-14 months of use.

 

Any recommendation as to what to do to improve this yield?

 

Thanks...

Link to comment
Share on other sites

Reformatting (zeroing) the three drives, two have now hard failed during format about 96%. Only one survived that trial. So two are coming in as RMA, and I'll try a couple more format on the sole survivor - I doubt it'll last much longer.

 

Some of the forum posts say it's user fault causing premature failures like this. It would be interesting to get a recommendation from the experts what to do differently.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...