[pwittman]

Hi All.

I hope that I'm posting in the correct place (let me know if I should pack up and leave). I've had good luck in the HD forum in the past, so I've decided to start my quest for salvation here.

So let me just line the whole thing out: Recently, a collegue of mine added a HD to our server - a 70 Gig SCSI Seagate Cheetah 10K.7 (ST373207LW), SCSI ID#8. This new drive makes the third on the chain -- sharing an on-board SCSI port (Ataptec AIC 7902 based Ultra320) with another Seagate Cheetah (35Gig, ST336607LW, ID#0) and a 17Gig IBM drive (DDYS T18350N, ID#2). There is a Seagate SCSI DAT Tape Drive plugged into a different on-board SCSI port. The Mobo is a Supermicro X5DA8 with Dual Xeon 2.4 GHz. The server is running Microsoft Server 2003.

The problem is this: When users on the network open files on the new drive (which is used solely as central file storage) -- and, subsequently, try to save them -- they will sometimes (with increasing frequency it seems) encounter extreme delays in the saving process. These delays often translate into the user's application (word, wordperfect, excel, etc.) locking up -- and sometimes (more importantly) the server itself crashes. The server crash manifests as a depressingly static grey screen.

The plot thickens: Because of all this, I have been doing some backing up of files on the drive in question to an external USB HD (Buffalo, 120Gig). One day, while doing some file transfers from the Buffalo drive back on to the Seagate Cheetah I got the dreaded blue screen. The FIRST time it said IRQL_NOT_LESS_OR_ EQUAL and the NEXT TWO times is said KERNAL_DATA_INPAGE_ERROR. :4-dontkno

Clearly this problem runs deep. And there is more I could tell you. For now, though, I'm just looking for sage advice on how to troubleshoot what's going on.

I've already run Seagate's SeaTools -- which, as I'm sure it want's to do, told me that my drive is fine.

Any thoughts? This madness cannot stand.

 

[linderman]

Welcome to TSF:

I dont see a problem with running 3 scsi drives on one chain ( i am doing exactly that now) however I have never tried an ID as high as eight (one ID past the controllers ID of Seven) I would say for starters change the drives ID to 3 or 4 or 5 etc you have a drive assigned 0 and you have a drive assigned 2 why not trying changing the jumper pins on the drive to something less than 7 ????

also what brand and model and size PSU are you running ?????? it better be one hellva power plant to light up all that stuff !!! wow

how about system temps, CPU temps and voltages
use sandra lite to determine and please post back
http://www.download.com/3000-2086-10018691.html

regards

joe

 

[crazijoe]

The plot thickens: Because of all this, I have been doing some backing up of files on the drive in question to an external USB HD (Buffalo, 120Gig). One day, while doing some file transfers from the Buffalo drive back on to the Seagate Cheetah I got the dreaded blue screen. The FIRST time it said IRQL_NOT_LESS_OR_ EQUAL and the NEXT TWO times is said KERNAL_DATA_INPAGE_ERROR.

Check your event logs for an error pertaining to this. Also check the device manager for any hardware confilicts.
I would have to quote linderman on your pick for the SCSI ID on the drive. I would use something lower than the controller ID.

 

[pwittman]

thanks guys,

as always, i greatly appreciate the help.

here is what i did last night -- which seems to have fixed the issue :grin: :
i moved the 17Gig IBM drive to the SCSI chain/adapter with the tapedrive. now i have two drives on each scsi chain. i also replaced the cable on the "bad" chain and updated the SCSI adapter driver from the manufacturer's website (which may or may not have added support for a greater number of SCSI peripherals). also, because i was seeing the "KERNAL_DATA_INPAGE_ERROR" those two times i decided to increase the size of my paging file (to twice the size of the RAM) and add a second paging file on one of the other physical drives.

truthfully, i did all this is more-or-less one swoop... and, consequently, lacked the scientific rigor to put my finger on exactly what was wrong. unfortunately i didn't have the luxury of too much down-time... so i kind of did it all at once.

either way. it works now. and i'm fairly convinced it will stay that way.

still, i think that the next chance i get to shut down the machine i'm going to take your advice and change the SCSI ID # to a lower number than the adapter. i'm sure it can't hurt, anyways.

also, you've raised my blood pressure a bit by asking about the PSU. it's a 420W Ablecom SP420-RP. is that enough juice? should i worry about that?

 

[linderman]

Greetings:

I am shocked you are able to run that load on a 420 watt rated (they never put out the wattage as labeled) PSU

I would at your convience upgrade to a 550 or 600 watt enermax or antec PSU

double check your mobo specs or ask within the forum for assistance / some severs mobo's require a version 2.03 complaint PSU (they need two 5 volt lines) as well as two 12V rails

SCSI components are expensive without figuring in the down time on a business / I guarantee you wont regret the upgrade!

also / use a monitoring utility to check your current system temps, cpu temps, hard drive temps etc its very good info to monitor / especailly with a business server !! its not just Music files your worried about !!!!

http://www.lavalys.hu/products/overview.php?pid=1&lang=en


regards

Joe