LinuxQuestions.org
View the Most Wanted LQ Wiki articles.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Hardware
User Name
Password
Linux - Hardware This forum is for Hardware issues.
Having trouble installing a piece of hardware? Want to know if that peripheral is compatible with Linux?

Notices



Reply
 
Search this Thread
Old 09-12-2008, 12:02 AM   #1
mcd
Member
 
Registered: Aug 2003
Location: Boulder, CO
Distribution: Slackware, RHEL, CentOS
Posts: 825

Rep: Reputation: 33
Either bad raid card or bad drives...


I could use some advice and brainstorming, right now I'm really frustrated and nothing quite makes sense.

I had a computer I built with an Intel motherboard, blah blah. I bought a RocketRaid 1640 (with 4 SATA ports) and attached 2 Seagate 500GB drives and 2 Maxtor 500Gb drives. I put them in a RAID 10 and for the most part all was well. Occasionally when I rebooted the RocketRaid bios utility would fail to detect one of the drives (never the same one), but I would just reboot and it would be fine.

But one day I rebooted and there was a drive that wasn't detected. Subsequent reboots didn't help. Leaving it over night didn't help. Unplugging and replugging things in didn't help. When I started moving drives around and attaching them to different ports I got very inconsistent results. I swapped SATA cables too. In the end I decided it was the RocketRaid card, since if I played around with the order and number of drives attached, I could detect all the drives (just not all 4 at the same time). So I fought with the manufacturer and got them to issue an RMA. The new card arrived and I plugged in all my drives.... and it didn't detect one.

So I sat down with a piece of paper and a pen and tried every combination of drives in every order. Eventually I decided that one of the Maxtors had failed. So I bought a replacement Hitachi 500Gb drive. When that arrived I put it in and WOOT! all four drives were detected. I chose to rebuild the array and went do dinner. The rebuild process failed halfway through. I tried it again, same result.

I tried using different combinations of the three original drives with the new Hitachi but the rebuild failed every time. I decided that since it was a RAID 10 I should be able to break it and still have one mirrored pair of drives - that was a BAD decision. Now there is no container detected. I have four drives with tons of data, and one drive with nothing on it. OK. So I put in the "good" Maxtor and the new Hitachi, created a RAID 1, and chose to duplicate data from the Maxtor to the Hitachi. It failed at 24% ("Failed to copy, press any key to continue". I did the same thing with one of the Seagates and the Hitachi - same error but at 27%.

What are the chances that I have 3 bad drives?? That seems extremely unlikely to me. Even two strains my credulity - even the four original drives are only about a year old, and the Hitachi is brand new. So maybe it's the RocketRaid; but I have the same problem - what are the chances that I got two in a row that are both duds??

I tried running Easus data recovery on the Maxtor (I've had great success with that product in the past) and was able to recover hundreds of thousands of files, but none of the file names were preserved - everything is FILE6401.mp3 or whatever. So that's sort of a last resort, not very appealing.

Alright, so if you've read all this, thank you! If you have any comments or experience with RocketRaid or RAIDs or whatever that you want to share, please speak up. I think my plan now will be to put each drive, one at a time, into a USB sled and attach them to my laptop. I'd like to start with the Hitachi, and perform some read/write tests as intensively as possible, and try to figure out whether the thing is good or bad. Does anyone know any good utilities for doing that? Or even suggestions for writing a script? Can I just write a series of numbers to the device and then go back and read them? I have very little experience to go on (although my scripting is pretty good). And then, what should I do with the original four drives? Are there any suggestions for trying to recover data? Any special linux command-line utilities or tricks that might help pull stuff off? Does anyone have any good websites or resources they can point me to?

Thanks!

McD

PS. I have an odd assortment of smaller drives that I can use to hold any data I manage to restore.
 
Old 09-12-2008, 10:28 AM   #2
salasi
Senior Member
 
Registered: Jul 2007
Location: Directly above centre of the earth, UK
Distribution: SuSE, plus some hopping
Posts: 3,919

Rep: Reputation: 779Reputation: 779Reputation: 779Reputation: 779Reputation: 779Reputation: 779Reputation: 779
Quote:
Originally Posted by mcd View Post
I bought a RocketRaid 1640 (with 4 SATA ports) and attached 2 Seagate 500GB drives and 2 Maxtor 500Gb drives. I put them in a RAID 10 and for the most part all was well. Occasionally when I rebooted the RocketRaid bios utility would fail to detect one of the drives (never the same one), but I would just reboot and it would be fine.
I'm not going to coment on the rest, but IMHO, this is where your problems started (and is often a problem with the cheap RAID solutions, <rant> and as for the people who think that a cheap raid array makes their data safe, so that they can take less care about backups, words fail me </rant>).

On start-up, if it is marginal whether drives join the raid array in time, a drive might be ignored by the raid firmware. this drive has an incomplete set of data, but (allegedly) that doesn't matter because you've got the full data set (indirectly) on the other disks.

And then it happens again, with a different drive, and now you haven't got all the data unless you've got all the drives and the redundancy that you thought you were getting doesn't exist. And this problem gradually gets worse over time, unless you are doing on the fly rebuilds, but you probably don't have monitoring tools in place to know that you have a problem, so why would you do that?

And then something 'real' goes wrong and you are in a world of trouble.

Sorry about the unhelpful rant, but if you buy an expensive, pro-spec, raid card, and SCSI drives (seeing the word Adaptec, for example would be one of the few things that would give me confidence, but confidence coupled with an empty bank account), you won't have this problem. I won't use cheap raid, because I've seen and heard of too many problems. The trouble is, I'm sure these thing get better over the years and I can't see what might happen to make me want to try raid to find out whether it has moved on far enough.

In the interim, I can only say that RAID makes backups more essential rather than less (and I even heard of one total something-head of a "professional" IT technician who claimed "We don't need to do backups, we've got raid" - this is incredible) when you take into account that it probably increases your tolerance to failure modes that you understand, and increases the susceptibility to ones that you don't.

Sorry that didn't do anything directly to help.
 
Old 09-12-2008, 11:59 AM   #3
mostlyharmless
Senior Member
 
Registered: Jan 2008
Distribution: Slackware -current (multilib) with kernel 3.16.2
Posts: 1,571
Blog Entries: 13

Rep: Reputation: 182Reputation: 182
On a hopefully more helpful note (though I sort of agree with the above), sometimes, even with modern motherboards, it helps if the RAID card is in a different slot. Just like you did with the drives, get a piece of paper and swap the cards around in the machine and see if the card is more reliable in another slot.

Or not, and give the thing up as a bad job.

For that matter, it might just not be seated well, or have dirty contacts. I had a problems with my cheap Promise card (I "Promise" I'll never buy one again) until I did extensive brain surgery on my computer.

BTW, I agree, backups are more critical with RAID, not less.

Last edited by mostlyharmless; 09-12-2008 at 11:59 AM. Reason: typo
 
Old 09-12-2008, 12:29 PM   #4
mcd
Member
 
Registered: Aug 2003
Location: Boulder, CO
Distribution: Slackware, RHEL, CentOS
Posts: 825

Original Poster
Rep: Reputation: 33
Well, this was a home computer housing personal data. If I can't recover any data it isn't anything more serious than an inconvenience and a serious disappointment. I didn't have the resources to back up almost a TB of data, unfortunately. The RAID was supposed to give me *some* protection against hardware failure, but I agree that backups would offer far more protection.

Anyway, to the previous poster, thanks for that bit of advice and I'll give it a shot.

McD
 
Old 09-12-2008, 07:26 PM   #5
Bruce Hill
HCL Maintainer
 
Registered: Jun 2003
Location: Tupelo, MS
Distribution: Gentoo
Posts: 6,926

Rep: Reputation: 124Reputation: 124
Something you might consider is using a cron job to do a
nightly backup. We do this for the computers on our LAN,
to the server. I'm presently trying software RAID in *nix
for the first time ever. As previously stated, my only
other RAID experience has been with proper RAID controllers
and SCSI drives. Those nightly backups provide some sane
amount of personal data, such as email and files of notes,
to be in place on the server.
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
mount: wrong fs type, bad option, bad superblock on /dev/hdc3 Lordandmaker Linux - Hardware 15 06-05-2011 09:55 AM
Bad mount of .mdf - "wrong fs type, bad option, bad superblock, on /dev/loop0" Maybe-not Linux - General 2 02-29-2008 02:30 PM
mount: wrong fs type, bad option, bad superblock on /dev/sda1 yekim Linux - Hardware 10 07-05-2007 12:28 PM
mount: wrong fs type, bad option, bad superblock on /dev/cdrom, Adil_uk Linux - Hardware 9 02-16-2005 06:50 PM
mount: wrong fs type, bad option, bad superblock on /dev/sdb1 pmoreira Linux - Hardware 5 02-26-2004 11:31 AM


All times are GMT -5. The time now is 02:17 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
identi.ca: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration