A chkdisk is NOT an appropriate mechanism to fix parity errors. It will only examine blocks where filesystem data exists, and then only a subset. You need to either run the software that came with the controller, or a program that reads all logical blocks and forces a media verify. Depending on exact make/model of controller, another issue is that the controller may not be configured to perform a XOR check on a read! Some software RAID does this, some LINUX kernels do this also.
Also, there are blocks on the logical RAID volume that are leftovers and not part of the filesystem. The physical blocks that map to them can also go bad, even though they are not part of the file system. If 2 blocks go bad, and you have a drive failure, then you are going to be in big trouble when it tries to rebuild and gets to the place where you have a drive failure, and an unreadable block on the surviving disk.
If for some reason you can only run run them when booted to the BIOS, you could try something like santools' smartmon-ux software and use the -verify command, which will perform a full volume verification on the logical device (i.e. \\.\PHYSICALDRIVEx), and forces the RAID engine to validate 100% of it. This can be done online also but it will obviously impact performance while it is running. Not all RAID controllers support this command, but most do.
Main Topics
Browse All Topics
by: Dusan_BajicPosted on 2009-11-06 at 06:42:04ID: 25759460
I have done it it successfully on external Dell raid5 SCSI storage, it took a lot of time but it fixed errors.