Raid Partition Mystery

[Ant2000]

We had to use a data recovery expert to get data from our server. He was nice and friendly and got the data within 2 hours.

Scenario: When we restarted Windows2003 server one partition of Raid 5 was wiped out. There are two partition on this raid. C and E drive. E drive had problem, it said its RAW, disk manager showed size 60gb, availabe 60gb. Under my computer, when double click on E drive it asked to format, it wasn't showing any info.

Action taken: This data recovery engineer also doesn't know what happened but he recovered all the data for us. He used a software, copied everything to his hard drive, formatted E drive(which had problem) and copied back the recovered files into it, and we were back online in 2 hours.

Now we need to know/find the reasons why just a partition would just go crazy. I asked few IT people and they also don't know the reasons. I am sure some popular people like, blakst, maysoft, doomer, shahij, hdspecialist, just to name a few here, would be able to give us some reasons why a partition would be wiped out.

Thank you in advance.

[rchadwick]

Maybe someone else has some insights (Especially with more details... More details about the RAID, Controller, did you do HW diags, etc), but personally I think it likely comes under the heading of Sh1t Happens.

If you're just concerned this doesn't happen again, do full HW tests on the computer. Also, back up like crazy for a while. You might also want to do some research on your specific hardware. Some chipsets, with certain driver versions, have been known to corrupt data, and do all kinds of other undesirable seemingly random things.

[quasimodo]

It could have been many things such as: app code gone wild on your server, unproper shutdown( power outage, etc..) It only takes a bit accidently writen to the wrong section of your hard drive for such problems. Often people also forget to defragment their Servers and think that just because it's a RAID set, there would exist no fragmentation.


Or a logical bad sector in your partition table perhaps.

[pepe]

Hello,

I think this problem has nothing to do with the raid, the FS of the 'E' partition got corrupted simply.
this could be caused by a problem in the OS kernel, system memory, overclocking (hopefully a server is not overclocked, but who knows...), power supply, bad IDE cable/connection...
The HDDs themselves can be excluded to be the source of the problem, since in such case the whole RAID array would have failed and u could not have accessed the 'C' partition either.

pepe

[shahij]

It's a File System corruption into the partition table. may be also error sector (called logical bad also)
Before formatting the drive, you could see the partition table, you could get exact report.

[N.C.]

Hello, Ant2000

The main problem is, the server doesn't have operating system, only have windoze....
This is generally a bad idea....

Additionally, you can support the badness with worst quality hardware, as pepe said.

I am always suggest to my all customers, replace the ps + hdd in every 2 year period on important places, even if have no problem at all!

Windows can corrupt own fs some time, i am no wondering about it.

Regards,
Janos

[Ant2000]

Thank you everyone for your help.

[BlackST]

On 2k3, probably the issue was a problem with raid. Often in botchered configs, where the main argument while choosing ir the co$t... Never rely on integrated raid. And, a common pc is not a server, just like a Toyota Corolla won't turn into a F1 if you put a FERRARI sticker on it.

[scratchy]

Another problem it could be is actually one that is often overlooked for this type of problem - SMART

If the Boot sector for this partition became bad (or a sector in a critical MFT record), then SMART will kill it and reallocate it with a nice and shiny new one and totally cag up the OS - but the drive is happier and feeling much better

You would need to check the SMART logs and G-Lilst of the drive.

It just might be!

<itch>