P5B-Deluxe ICH8R RAID is unstable!

P5B-Deluxe ICH8R RAID

ICH8R RAID is unstable!

I set them machine up on Friday and have been installing software and gaming some. The array has had to re-build itself 5 or 6 times so far!

I’ve been building PC’s as a hobby and as a side profession since 1989, I’m not new to any of this. For $240 I hoped I’d be spared buying a $200 RAID controller but apparently not!

I can see it blowing up if I were overclocking crazy but I’m not. Its blown up even while not overclocking. Once just changing something in the bios, saved then BAM! 3 hours later I can actually use the machine again.

Hard drives are new Maxtor Diamond 11 400GB models. RAID-1

Core2 Duo E6400 CPU

2 GB G.Skills DDR2800 PC6400pro

ATI X1900XTX video card

WinXP Pro, CAT 6.9 drivers, latest Asus bios and updated drivers from everyone.

Each time this happens, it takes 2 to 3 hours to re-build. I don’t have the time to waste, maybe I should start sending Asus a bill for my time lost.

Before I go spend as much on a raid card as I did on this motherboard, anyone have any suggestions?

never had this problem, also I been on these and XS forums reading about the P5b since it came into existence and this is the first time I heard this. Either your OC is unstable or you got bad ram or bad drives

Rattle said:

never had this problem, also I been on these and XS forums since the P5b came into existence and this is the first time I heard this. Either your OC is unstable or you got bad ram or bad drives

Click to expand...

OC at 8 X 440 Prime95 for days while folding, 3D Marks in loops for hours upon hours, TMPEG rendering video, Adobe Premiere Pro video editing and so on without a single crash.

I OC my video card till it locks up. System reboots and the array is failed. This happened 3 times. Why? Twice the system wasn’t even OC’ed, so overclocking the video card caused the RAID to become unstable?
What’s obvious is that it occurs after the system reboots abruptly. This was the same issues I had with the ICH5’s on several boards I used in my office. I bought PCI raid controllers because the time lost and chance of loosing data was not acceptable. The power would occasionally blink out.

I shut the PC down, unplug both hard drives so I could bleed air out of the system. Shut it down, plug the drives back in and the array has failed? I think not!
At home the problem got so annoying with the ICH5 that I bought a Promise T2300. I never once had the array fail again, regardless of how hard I crashed the system OCing.

How many people are OCing and running RAID-1 with the ICH8? Most? Few?

boead said:

I OC my video card till it locks up. System reboots and the array is failed. This happened 3 times.
Why? Twice the system wasn’t even OC’ed, so overclocking the video card caused the RAID to become unstable?
What’s obvious is that it occurs after the system reboots abruptly.

Click to expand...

Well, look at the highlighted text I made, it seems like you've summarized it pretty well your self !

See the pattern ? Locked up then the raid was rebuilding again, of course your understanding of this problem that when it was locked up while torturing your GFX card, it was not supposed to affect the raid drives right ?

Just to remind you even there wasn't any disk access at the time it locked up, put aside Intel ICH raid driver, the Window OS it self will detect it and know there was a dirty shutdown and normally it will do some drive self checking on next boot up. The ICHR driver just ride on that status/mechanism and do what it supposed to do in ensuring data integrity.

Now, actually it was not rebuilding, it was verifying and repairing it if necessary, since it is Raid 1 and they must have an identical copies on each drive.

How do you know "EGGxactly" if those drives were indentical when the time your puter went kaput ? Heck, no body can proof that either !




See this familiar pic above ? When it starting to do the checking, you can stop it by right click the Raid 1 volume then choose to stop the reparing/verifying process. Of course you're aware what are the consequences by stopping that process right ?

Now, a bit details, assuming you're using the latest driver version 6.1.0.1002 and just read below notes on each revisions :

Here -> Intel® Matrix Storage Manager Production Version 6.1.0.1002 Release Notes

Quote on New Features Baseline 6.1 :

- Verify and Repair feature
- Automatically checks for data inconsistencies upon reboot after a dirty shutdown for RAID1/RAID5/RAID10 volumes and repairs them (RAID1/RAID5/RAID10 only)

Click to expand...

While lower version like below doesn't have it, check it out your self

Here -> Intel® Matrix Storage Manager Production Version 6.0.0.1022 Release Notes

See, if you feel it is "problem" instead of a "feature", all you need is just downgrade the driver from 6.1.0.1002 to 6.0.0.1022.

boead said:

I bought PCI raid controllers because the time lost and chance of loosing data was not acceptable. The power would occasionally blink out.

How many people are OCing and running RAID-1 with the ICH8? Most? Few?

Click to expand...

Since you mentioned that you ain't noob and understand what is the purpose goal of raid 1, just to remind you that actually its really helping you on your own expectation in "loosing data was not acceptable" by ensuring your important stuffs in there wont get corrupted, hey, isn't that nice thing to have ?

About PCI raid, forget bout it if it is not high end above > $200/card with battery backup cache and with inboard "true" processing instead of cheapo XOR-ing card, your ICH8R beats any of those PCI raid card like helll since they're bandwidth limited by PCI. Btw, PCI is not equal to PCI-X or PCI-E.

How many you ask ? A LOT ! Just click my sig's smiley to see how many, and actually you might interest to do some "bullet ducking" tricks like in "matrix" with your current investment, just read it through and hoping you'll take the red pill !

Anway if there are many problems on this ICHxR raid, just hang on a bit on storage forum, trust me, you will hear a lot of loud noise there !



@ MOD : I believe this thread should belong into "Storage" forum don't you think ?


-

boead said:

The array has had to re-build itself 5 or 6 times so far!

Click to expand...

Just so you don't feel lonely... I have had the same 'problem' with my ICHR7 setup on the P5W-DH... but only when I appear to have reached the overclocking maximum. I haven't persisted to work out exactly what causes it, but simply assumed that I was clocking too high. I can do a stable superpi run at 415fsb but if I push to 418 or 420, she will stall part-way through, lock the machine up and generally reboot, and that's when the array (only ever the Raid 5 thankfully) checks itself for errors. It actually found about 5 once and corrected them, but other than a time-consumer, it is not a problem.

For the record, my X1900XT is at stock setings, and memory is either 1:1 or 4:5 I think it is. I actually wondered if the board was overclocking the drive-controller somehow, but doubt that.

Anyways... said all that, just to say... don't feel so lucky... you are not alone with raid degradation (not that it is really... just being safe I guess)

My 2c worth.

Thanks guys, I don’t feel so lonely anymore. Fritzman, BIG fish you got there!

I got to say, the performance of these drives and ICH8 is the best I’ve had yet.

My last rig lasted the longest of any of my Frankensteins, over 2 years with just a few upgrades; P4c > P4e, XMS 2800 to OCZ 3200, Western Digital 200GB ATA-100’s > Maxtor 300GB Diamond10 SATA-II’s and a Promise SATA-II PCI card that supported NCQ. The performance gain from the onboard ICH5 (Abit IC7-G) to the Promise card was not nearly as significant as it is to this ICH8.
I built this machine and another for my office which is a P5B/E6300 using the very same Promise T2300 RAID card and a set of Maxtor 300GB drives. The difference is actually more then significant, the ICH8 is much faster! I really don’t want to blow a minimum of $140 on a HotPoint PCIe4X raid card knowing the differences are null. The Promise cards are just way to expensive right now for anything fast but as what’s been said, they have serious onboard processing, memory and the PCIe allows bandwidth beyond its needs.

Bing; YES Verifing and Repairing!! Ugggg. Didn’t know it could be stopped, I looked for it but it wasn’t obvious – LOL!
Very helpful, thanks again. I didn’t realize this was a ‘feature’ and not a ‘problem’ but some say the glass is half empty, I say it’s not quite full.

Still, canceling the Verifing and Repairing task isn’t smart so I still have to live with it! However, I can NOW stop it and start it at a later time when I’m not using the PC.
Reminder to self: Start reading the fine print and release notes!


Question; Volume write back cache is NOT enabled. Should I enable it?

boead said:

Question; Volume write back cache is NOT enabled. Should I enable it?

Click to expand...

LOL, that is your biggest loss if you don't enable it !

Now, the fun part, don't enable it yet, start reboot your windows and clock them how long it takes to boot and enable it and see the different your self !!

Well, a Hd benchmark program like HDTach before and after also good to see the difference, just prepare to hold your jaw before it drops ! LOL

Btw, if you go Matrix Raid, it will even lot better in performance compared to your plain raid.

Bing -

Re: Post #4

Excellent explanation.

JMHO,

~ Strat ~

Stratcat said:

Bing -

Re: Post #4

Excellent explanation.

JMHO,

~ Strat ~

Click to expand...

YUP give that guy another Gold star just saved me 20 hours of reading when i upgrade or crash and dont know what is going on.
have run raid for 5 years, and never would have thought they would do something like that. hmm feature, ok i guess?

boead said:

ICH8R RAID is unstable!

I set them machine up on Friday and have been installing software and gaming some. The array has had to re-build itself 5 or 6 times so far!

I’ve been building PC’s as a hobby and as a side profession since 1989, I’m not new to any of this. For $240 I hoped I’d be spared buying a $200 RAID controller but apparently not!

I can see it blowing up if I were overclocking crazy but I’m not. Its blown up even while not overclocking. Once just changing something in the bios, saved then BAM! 3 hours later I can actually use the machine again.

Hard drives are new Maxtor Diamond 11 400GB models. RAID-1

Core2 Duo E6400 CPU

2 GB G.Skills DDR2800 PC6400pro

ATI X1900XTX video card

WinXP Pro, CAT 6.9 drivers, latest Asus bios and updated drivers from everyone.

Each time this happens, it takes 2 to 3 hours to re-build. I don’t have the time to waste, maybe I should start sending Asus a bill for my time lost.

Before I go spend as much on a raid card as I did on this motherboard, anyone have any suggestions?

Click to expand...

Your OC is unstable, I bet ya. Either memory, CPU, PCI-E.
I have been running ICH RAID since the days of P5AD2-E. The ONLY time arrays EVER rebuilt themselves was when I was pushing the max of my system OC.

RangerXLT8 said:

Your OC is unstable, I bet ya. Either memory, CPU, PCI-E.
I have been running ICH RAID since the days of P5AD2-E. The ONLY time arrays EVER rebuilt themselves was when I was pushing the max of my system OC.

Click to expand...

Thanks Ranger. Bing answered the question. Its not unstable, Intel is doing me a favor.

Been beating the snot out of this rig and I’m really impressed with its stability and performance. I can get the FSB up some more and will do so in time but I’m not into the numbers game. An extra few MHz isn’t going to make or break anything. I usually throttle down a few notches from the top when I’m done and 440x7 will likely be where it stays.

Just got back from a film shoot and will be doing some editing tonight. Adobe Premier Pro just LOVES this processor.

Do you mean like this... ?



(PS... it's been that way since I set the system up)

Thanks markymoo, no need to shout. Bing answered the question. The array is only Verifying and repairing as is needed. Apparently it’s a new feature built into the newest BIOS that tells the array to do this each time it’s abruptly shut down. It can be canceled, stopped and started at ant time. Scroll up a little and read the thread.

PCI locks are nice to have. My Abit IC7-G had it. I wish we had it years ago; it would have been very helpful.

I had an Asus P3V-133 with the VIA 133A chipset years ago. Everyone was using the famous Intel 440BX chipset boards yet my P3B ran a 160Mhz FSB. I had a P3 550e (5.5 X 100) at 880Mhz with Muskin PC133 rev2 memory. Bandwidth was amazing for the time. I had that rig with a Voodoo3, the original GForce DDR and finally with the first Radeon. The P3B had a divider for 133Mhz so 160Mhz wasn’t too far off but the intel 440BX didn’t and you had to run the PCI way out of spec.

boead said:

Thanks markymoo, no need to shout. Bing answered the question. The array is only Verifying and repairing as is needed. Apparently it’s a new feature built into the newest BIOS that tells the array to do this each time it’s abruptly shut down. It can be canceled, stopped and started at ant time. Scroll up a little and read the thread.

Click to expand...

That makes almost no sense. A properly functioning array will never get errors via regular use. There is obviously some corruption going on causing the array to rebuild.

I agree... when I'm approaching the upper limits in fsb, occasionally it will restart and the array is fine so it doesn't self-check. When I get high enough however, she locks up and I need the reset button and when it restarts, it self-checks... sometimes finds no prob's, others finds 3 or so erros and corrects them.

I am glad it does.

CORRECTIONS

another addition, you know the chip for that P965 thing does get rather hot. and asus put that CAP thing over it for thier LOGO junk, that Slightly reduces the cooling of that chip.
the heat tube to the other heat sync is not really vastly effective at motivating the extra heat to the other sink, and is also slightly dependant on specific temps, and direction of the tube.

if that ASUS cap was not on that heat sink it sure would improve its capacity to cool, BUT removing it will move the thermal connection made (thermal goop) and void your warrenty.

because i didnt like it on there, i had to tear the whole thing out , re goop it, and plop it back in, with these notes:
i slightly bent the heat tube, but LUCKILY not enough to squish or corner it.
the thing must be flattened and straightened or the tube will cause a side lifting, reducing contac
i busted the dang pins , because i didnt want to tear it out of the whole computer, so i had to replace the pins.

basically what i am saying, is to very slightly improve the cooling capacity for the P965 chipset thing, i had to redo the whole thing.

not saying that is your problem, but if you offset bumped or altered that thing when shoving in some mondo cpu cooler or something, at the least the thermal connection could be reduced.

the Bottom chip thing ICH8, with the little heat sync still gets hotter than the p695 one, especially with it hidden in the hole there in my computer. the bottom one was at 130F when the top one was at about 113F. so in my opinion the ICH8 chip is NOT being cooled enough, i intend to put a larger passive sinc on it.

Just thoughts.
i dont know enough about it, but i am trying to make a quiet slightly fast computer, and i have a cheap thermal probe thing, because often we Miss some of the Hotest items in the computer, when overcooling the ones that fall under the hype and sales.
ex: on my server board, the voltage regulators were hotter than the cpu, but nobody mentioned that anywhere on the web.

sorry i corrected the post, got my chips mixed up, i should just say the Hot one , or the SATA one instead of trying to use the coded number junk

above post had to be corrected.

So far I have been up to 510 fsb with 2 raptors in raid 0,a ide seagate off the j micron controller and a ata 66 card for my dvd-r drive in the pci slot.Haven't had a single fail of my array,the title of your post is misleading.

RangerXLT8 said:

That makes almost no sense. A properly functioning array will never get errors via regular use. There is obviously some corruption going on causing the array to rebuild.

Click to expand...

You’re not reading the thread. We already established that ITS NOT rebuilding the array but rather checking its integrity after a hard crash. It can be stopped and ignored entirely, I have yet to have it find any errors.
Again, read the entire thread, its short enough.

But its of no consequence, I’m doing fine with it. Happy at 3560Mhz. Haven’t had a single crash and I’ve been video editing in Adobe Premier Pro for hours upon hours.
Nice and fast.


Game performance is great too. Can’t wait to Crossfire this rig with another X1900 in a few months.