Odd Hard Drive Failure

[Diego Garcia]

This isn't a Sage problem (at least, I sure >>hope<< it's not a Sage problem ), but I thought I'd post it here because of the high level of experience with heavily-used disk drives. Besides, the problem is on my Sage server .

A couple of months ago I installed a Western Digital 1 Terabyte "green" drive (i.e., it's designed to spin down to 4500 RPM when not being accessed).

A few days ago the drive disappeared from within Windows Explorer (I'm running Vista 32 SP2 with all current patches), apparently while Sage was recording to it, as Sage also crashed. I popped over to the disk administration console under Administrative Tools and saw that the disk was no longer listed. So I told the disk administrator to rescan the drives...which locked up my system so tight I had to shut it down and power it back up in Safe Mode. The "disappeared" drive was then visible, and running chkdsk on it allowed me to see and use it again in normal Windows mode.

However, the problem recurred last night. This time I checked the error logs, and there were a number of Event ID 51 warnings ("An error was detected on device \Device\Harddisk2\DR2 during a paging operation"), followed by an Event ID 57 warning ("The system failed to flush data to the transaction log. Corruption may occur"), all occurring right before the drive vanished, so far as I can tell. This second failure apparently did not involve Sage, as it was recording shows to the second, smaller hard disk in the system while I was doing some spelunking.

Western Digital's diagnostic software found no problems with the drive, not even with the three hour long intensive testing regimen. Google shows a number of reports of the identical problem, not all involving WD drives, but I've yet to find a consensus on what caused the problem. One fix involved changing the SATA data cable, which I've done.

Has anyone else run into this problem? If so, what solutions did you find?

[zoundz]

If you google the event code 51 you will find a lot of people having a lot of different problems that cause that event. In my case, I had a lot of problems with 3 different 500GB SATA II drives from both Seagate and WD. Every once in a while I'd see a sequence of event 51's and then the system would lock up. Tried a bunch of stuff to fix it. I had a 200GB SATA II Seagate that had no probs whatsoever in the same context. I tried getting rid of the 64K blocks on the 500's. I tried different cables. I tried a new PS in the box with built in SATA power cables. All to no avail. By the time all was said and done, the only conclusion was that the SATA chipset in my MB just has trouble with 500GB SATA II drives. The final solution for me was to go back to PATA drives on that machine. No probs since. Those same SATA II drives are totally happy in my Thecus 4100Pro NAS. I don't know what MB you have, but my ECS 755A2 with the SIS964 SATA controller just couldn't do it. The SATA II drives are supposed to be backward compatible with the SATA 1 chipsets, but in this case, the 500 gig drives just wouldn't work.

All this may or may not be relevant to your situation, but I thought you might feel better knowing you aren't alone <g>

[MattHelm]

My dad's server (see #2 in sig) seems to have something like this going on at times. I read somewhere that it might be a lack of memory problem, and so changed his ram to 2G from 1G. So far, all has been fine, but his only happened once in a great while, and did it with another drive I had in the system, which seems to support the memory or MB problem. But, the MB is the same as mine (#1 in sig), and I've never had this problem!

Hope this helps you find the problem.

[Diego Garcia]

Thanks to both of you for the quick responses. One of the reports I've come across is that the nVidia chipset (which I believe my MB uses) may be implicated as having some problems dealing with SATA drives. I sure hope that's not my problem, as I'd rather not rebuild the server.

Does anyone else have experience with this problem?

[waltraud]

Hi,

Event ID 51 Lost delayed write basically means that Windows tried to flush its memory buffer to disk but wasn't able to do so. This usually happens when a storage device was removed but not properly dismounted beforehand ("Safely rmove hardware" Icon).

In your case I'd say the disk is randomly disconnecting from the bus, so Windows is loosing access to the disk. This can have multiple causes

- The controller on the disk is faulty
- Your MBs Sata Controller is faulty
- Your SATA cable is faulty
- There is a compatibility issue with the disk and the SATA controller

This is what i'd try:

- If the drive is a SATA300 drive try setting it to SATA150. There should be a jumper on the drive which allows you to chnage the SATA Mode.
- Update the BIOS of your MB and the SATA controller drivers
- If possible try the drive in another PC
- Replace the drive....

Martin