I have four WD2000JB drives connected to a Promise FastTrak TX4000 controller in a RAID 0+1 array.
I suspect that one or more of the drives is having an intermittent problem, but Promise doesn’t provide any diagnostics that can test for (or even report!) intermittent drive problems while the disks are connected to their controller. Their only recommendation is to disconnect each drive from their controller and attach it directly to an IDE port and then run “the drive manufacturers diagnostics”.
OK, I can do that. But how comprehensive are the Data Lifeguard (for DOS) diagnostics if I only run the “read-only” (non-destructive) tests?
And if I need to run “destructive” (ie, writing to the disk) diagnostics, how do I do a COMPLETE BIT-for-BIT backup of the disk… including all the array config info that the Promise controller writes (somewhere) on the disks? And of course, how would I restore it after the destructive tests?
(BTW, I’ve already contacted Promise, and they are not much help).
This is not tech support, but I do know these products, and more than likely, you are having problems with them dropping out of RAID. That’s because the JB drives are not RAID drives. If they go into error correcting mode, as is common among desktop drives, they will drop out of the array. For true RAID configuration, you need to use our RE drives.
But the symptoms I’m seeing are lockups and intermittent extremely sluggish performance, not loss of members of the array. So perhaps it is (as you say) the drives going into error recovery, but if so, then the RAID controller must tolerate that, 'cause the drives don’t drop out of the array.
In any case, I still think I should run the WD diagnostics against each of the drives. And to do that (safely) I still need to know how to do a complete backup of the drive…
I’m sorry, but that’s not true. A true RAID array won’t tolerate more than about 7 seconds of error recovery before the drives start dropping out of the array. That’s what a RAID Edition drive is for: it won’t do more than 7 seconds of error recovery. Some desktops can go as long as 2 minutes in error recovery. Check the link down below.
Unfortunately, DLG Diagnostics does not read RAID arrays. You would have to read each drive individually, and with a RAID 0+1, you could lose all your data doing that. The way it sounds, though, I would get the data off the array and check the drives. Something is definitely off. Can Windows do error checking and recovery on a RAID array? If so, I’d try that first.
What is the difference between Desktop edition and RAID (Enterprise) edition hard drives?
