Hiya all.

Can you take a look at the screen and tell me if this is normal for drives of this era.....



This is in a Dell SC440 server with a SAS5 i/r controller with mirrored drives

The drives are a WDC WD2500JS-75N and a WDC WD2500JS-22N

I enabled disk casheing on the controller and get slightly better results......



Still the system is painfully slow

Newer better drives needed? Any other ideas?
Is there any software that can check the SMART attribs of each drive while still in raid?

These drives are ancient history. If that server holds critical useful info, I suggest you replace drives ASAP.

_________________
http://www.northwind.gr
SandForce SSD Recovery
Ransomware Reverse Engineering - NoMoreRansom! partners

@saggy,

+1 for northwind's comment

Seeing as the "drive" in both graphs seems to be the RAID device (labeled "Virtual Disk" by HD Tune), it doesn't tell you all the necessary detail IMHO. Certainly there are unexpected dips in the throughput (and I'd be worried by those graphs), but we can't tell if that is due to the one disk in the mirror, the other disk in the mirror, or something else.

On the RAID systems I work with, we have vendor-specific commands - one of which tells the RAID controller to send the I/O to a specific underlying physical drive for testing purposes (e.g. I can read from just one disk in a mirror or stripe or RAID5 etc.). Without having that detail on the system you have, then to test whether one (or both) drives is having issues reading some areas (which is one cause of dips in throughput), you may need to do things like test the drives individually on a non-RAID controller.

Regarding your question: The SMART attributes may or may not tell you something useful - IIRC the Dell-supplied utils with that controller can read the SMART data from the individual disks (but it's been a long long time since I touched one of these cards, so I might be misremembering). I believe this is "really" an LSI card - if so, the manufacturer's utility is LSIutil under *nix, but I don't know whether they supply an equivalent Windows utility. You could try the Windows port of smartmontools, to see what that can tell you...

Cheers Both of you.





Pretty conclusive really. Cheers to both of you.

I agree the 22N drive looks more sick in that data, and is probably the priority for replacement - but the 75N drive has been running for much longer than the 22N (e.g. if it's running 24x7 then it seems to have been running for approx 5.5 years?), so personally I'd follow the comment from northwind with that drive too.

Good luck & glad you were able to find some useful evidence.

Right......

Ive got 2 new drives but the raid card appears to be a bit of a pain.

The process I have in my head is to image the better out of the 2 existing drives, to file, using raw copy tool.
Then extract this to one of the new ones.
Put the new larger one with the data on in as the boot back in the server.
Then install the other drive and let it rebuilt itself.

This is the method I used once before with success - but is there an easyier way than this?

Cheers

These drives PCB are ticking time-bombs to, they die a looooooot