The Raid stripping is going bad and we are getting I/O errors on systems. We have had 14 in the last year. We do have a lot of servers in the field (hundreds). Once you stripping goes bad basically your raid will fail. We are not getting defunc drives, but our Seq. databases are getting corrupt. I thought there might be a software program to check your parity on the drives for something.
Maybe stripping goes bad over time? but I would think there should be a way to check that.
Jeff
On the adapters I use, Consistancy check, verifies the array data, including the parity info.
If an array is properly setup, with a hot spare, an array loss is rare. With over 20 raid5 arrays at different clients, in five over years, no losses.
With a few Dell servers, a bad firmware release caused repeated false drive failures. One array was destroyed due to a drive firmware update using Seagate's Enterprise drive utility.
In all the years I been using raid setups I have never had I/O errors unless.....
firmware on the array adapter is buggy
Firmware on drives is causing errors
Drives on the array do not match as far as firmware revision; Drives of different firmware revisions often work quite well together,other times they don't.
If your drives have a firmware revision of version 1 and the manufacturer is up to version 5, more than a few bugs were fixed
Scsi setup, termination or cable problems
Heat can cause a drive to have I/O errors.
Though rare.....
Firmware on the array adpter is easily upgradeable (and downgradeable in case of a problem). Still, upgrading firware on a raid can cause serious problems.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.