What does it mean "One or more logical drives contain a bad stripe: controller 1" (server RAIDmanager)

It means that your logical drive has an error.
A logical drive is a drive combined out of X physical drives.

In your case it means that one of you 4 disks seems to be bad.

If you are using Hp Server then try checking in the SML or ADU so that we can identify if all the disks are having issue or if any drive is faulty and this depends on the Raid being implemented

we can also look in IML to confirm if the controller is fine or not

Thank you,

i'm using ibm servsr "IBM Server xseries 226 "

How do I know which of the disks is damaged?

You cannot tell which disk is the problem. A bad stripe on a logical drive generally means that all disks have readable data but that parity doesn't match data. This can occur with RAID 5 if you have a power-out and the cache battery goes flat before it can write the contents to the disk. The error is most likely found when the controller is doing background parity checking which most of them do when they are idle, since the data and parity are inconsistant the controller doesn't know what to do to repair the problem. Only way out of the problem is to backup what you can and restore after re-initialising the array.

Some controllers have a re-sync function, don't know if yours has it though

i try to backup with acronis and get "faild to read data from the disk ".

I think that  one of the disks has a bad sectore.
;(

If one of the disks had a bad sector the controller would get the data off the other disks using parity so that can not cause the problem. You wouldn't have a problem if one of the disks had a bad sector. The controller can't reconstruct the data from parity with a bad stripe because it doesn't know which drive has faulty data. It therefore returns a read error to the OS. It is not a disk fault, it is a data mismatch.

Acronis is an imaging product, it will stop as soon as it reaches the read error. You have to do a file-by-file backup, you should be able to get all but one or two files off it, you might even get the whole lot since the bad stripe may not have any data on it. If the file-by-file backup fails when it reaches a bad file rather than carry on with the rest you have to delete that file and then try again.

There may be an option to re-sync but it would require manual intervention to tell which of the disks had the mis-matched data on it. RAID reconstructor from runtime.org may be able to deal with the problem but you still would have no way of knowing which version of the file is bad.

If you imaged all 4 disks seperately with RAID reconstructor (you would have to take them off the RAID controller and put on a non-RAID one for this) you could generate 4 recoveries, each with a disk missing, and compare them.

Taking one of the disks out and shoving it back in again will probably cure the bad stripe *but* you would not know if the data was valid or not. You would be randomly selecting which disk had the mismatching stripe element on it.

RAID level 2 would cope with this but nobody uses RAID 2 any more.

You could even ignore the error and continue to use it, the controller will give a read error whenever the bad stripe is encountered and the OS may automatically map it into its bad sector list (it's not a physical bad sector but to the OS it looks like one). chkdsk, fsck etc can probably map the blocks out for you but you may not be able to do an image backup even then.

Thank you,

i try to copy file by file and Unfortunately the data base of mail server not pass .