gerpaqit
asked on
Disk Failure
hardrive failed on my dell server today. it was buitl using a hardware mirror with two SATA drives.
The server still boots but all files have reverted to being no recenth than March 2010 which means i'm missing 9 months woth of work.
The server is a Poweredge sc430
The server still boots but all files have reverted to being no recenth than March 2010 which means i'm missing 9 months woth of work.
The server is a Poweredge sc430
Brian B
A hardware RAID1 should exactly a mirror. The controller maintains two drives as exact copies of each other. Unless you mirror failed back in March, I don't think this could happen. You might want to go over any logs you have from your Dell configuration manager to figure that out. If the data is critical and/or not backed up I'd suggest getting on with Dell support to find out exactly what happened before you accidentally write over the newer data.
If you were running a hardware Mirror on the 2 drives, from what you have stated one of the drives may have fallen out of the array 9 months ago but did not physically fail. Now it appears that the drive that was online and running, has failed, and you are booting to the 9 month old drive.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
I agree with the above. Your Array may have fallen apart the 9 months ago.
Check your backups for the data for sure!
Check your backups for the data for sure!
what happened is that the mirror was broken since march. sorry the data is gone. take the other drive to a data recovery firm with $1000 or so
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Go buy SpinRite 6.0 for $89 and run it against the drive that just failed. You can run it on another machine. There's no risk to your data unless the head crashed and you have bits of metal flying around inside the drive. If it doesn't get the drive running again, you can ask them for a refund.
To confirm if your mirror broke in March, check the event logs on the server. Do they go back like they should, or is there a giant gap in the dates?
How are your backups?
http://www.grc.com/sr/spinrite.htm
To confirm if your mirror broke in March, check the event logs on the server. Do they go back like they should, or is there a giant gap in the dates?
How are your backups?
http://www.grc.com/sr/spinrite.htm
No, there is extreme risk to a HDD in running spinrite. The controller dropped the HDD for a reason. The disk failed to respond in the appropriate amount of time. So at this point, you are on borrowed time. Every moment the HDD is powered on could be the last. The most stressful thing you can do to a HDD is pound it with I/O. (By the way, spinrite does nothing more than read, reread, reread disk blocks in an attempt to get data back. It is not unusual for it to take days on a large drive).
If you have NO plans to go to a data recovery firm to get the data, then before you run spinrite, I suggest trying to first try to clone the HDD and skip over the bad blocks. You will need another disk anyway, so might as well do that first and copy to the new disk.
By the way, making a binary copy of the new disk does not mean that the new disk will work in that controller, as the metadata will not match the serial number of the HDD, so you will still have to deal with stripping off the metadata using a copy with offset and binary editor to find the location of logical block #0. (Or if you are not up to using a binary editor to find the end of the metadata, then buy a product such as runtime.org's raid reconstructor. )
But the takeaway is that spinrite is NOT safe, so if the data is worth a lot of money to you, then consider even powering up the HDD is a risk, and you should take it to a recovery lab.
If you have NO plans to go to a data recovery firm to get the data, then before you run spinrite, I suggest trying to first try to clone the HDD and skip over the bad blocks. You will need another disk anyway, so might as well do that first and copy to the new disk.
By the way, making a binary copy of the new disk does not mean that the new disk will work in that controller, as the metadata will not match the serial number of the HDD, so you will still have to deal with stripping off the metadata using a copy with offset and binary editor to find the location of logical block #0. (Or if you are not up to using a binary editor to find the end of the metadata, then buy a product such as runtime.org's raid reconstructor. )
But the takeaway is that spinrite is NOT safe, so if the data is worth a lot of money to you, then consider even powering up the HDD is a risk, and you should take it to a recovery lab.
ASKER
Thanks for the input guys. Removed the bad drive and shipped it off to a data recovery specialist today. So fingers crossed.
Did not run any recovery on it myself but can hear the drive clicking so don't hold out much hope for a recovery.
Did not run any recovery on it myself but can hear the drive clicking so don't hold out much hope for a recovery.