dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
407
share rss forum feed


Inssomniak
The Glitch
Premium
join:2005-04-06
Cayuga, ON
kudos:2

Raid 5 troubles

Ok as my luck would have it, I have a 3 drive raid 5, have 2 drives fail. well thats not good.

So Drive 1 and the parity drive are reporting as failed, they are connected to a PERC 6i controller.

Drive 1 still seems to be readable but the dell server wont allow it.

Is there a possibility if the drive 1 "ghosts" or images to another drive of the same size that the controller might like it? Or does it go by serial numbers, etc ?
--
OptionsDSL Wireless Internet
»www.optionsdsl.ca

Chrno

join:2003-12-11
Chances are, the data, even if drive 1 is readable, is done. In raid 5, data is weaved across the 3 disks in stripes (or if you prefer to call it blocks). Stripe sizes will vary from configuration to configuration but some common sizes include 32K, 64K, 128K, 256K and so on. When data is written to the disks, the data is divided into any of the above chunk sizes (depending on configuration) and written the array. Say if your file is 32K, it will get divided into 2 16K chunks. 1 chuck is written onto drive 1 while the other onto drive 2. Drive 3 will contain the XOR parity (the 3 make up 1 stripe). With drive 1 readable, you only have 16K of what's suppose to be a 32K file which is completely useless.

What you can do at this point is bring the drives in for data recovery. They will try and work an image out of all 3 drives and attempt to reconstruct the array. If your data is important, avoid writing to any of the drives.


Jan Janowski
Premium
join:2000-06-18
Skokie, IL
Reviews:
·Comcast
reply to Inssomniak
Correct me if I'm wrong, but you'd need a Raid 6 in order to survive 2 drives failing....

Happened to me once..... I had one fail on a Raid 5, and I pulled the one of the remaining good drives, and inserted a fresh drive there..... OOPS!!!

I learned to be more particular looking at the LED's on the array after that... And to show more 'respect' for "Hot Swap" array's.....

I was incredibly lucky.... I had a program "Allway Sync" that made a copy of the raid onto an older, smaller, much slower, raid, and it had all but that day's data backed up....

So after resurrecting the raid, I fired up "Allway Sync" and restored that raid.... Got me to where I was the end of the day the day before....

Needless to say.... you should always have an identical spare drive available for these 'situations'... I'd go so far as to say same MFG, Model, and Firmware.... Just to cover your butt!! !
--
Looking for 1939 Indian Motocycle


Inssomniak
The Glitch
Premium
join:2005-04-06
Cayuga, ON
kudos:2
Wouldnt you have it, the RAID array restarted on 2 of the 3 drives, One is "good", other is "rebuilding" and 3rd is "predictive failure", so we are using the opportunity to GET BACKUPS !!

Production system in a DC, for 4 satellite credit unions, voip, exchange, big mess lemme tell ya.
--
OptionsDSL Wireless Internet
»www.optionsdsl.ca