Ratdude's Main rig V2

[Topcat]

Re: Ratdude's Main rig V2

PAE issues? I thought this was the system you had running it. Corrupt HDD other wise screams out write caching corruption which could be a bad HDD itself or a controller issue. I used to have an adaptec SCSI controller that would do this exact thing. I finally got annoyed with it and shitcanned the controller, which fixed it. This was back in the U2W days.

[ratdude747]

Re: Ratdude's Main rig V2

And she died again, same symptoms. Crash, reboot, RAID array failed. AFAIK the BBU is OK.

I've tried the following:

-Re-recapped the PSU
-Memtest86 passed (so not a RAM issue?)
-Swapped BBU batteries (Or at least I think I did)
-Swapped RAID Cards and BBU controllers
-Swapped HDDs and SATA cables.

So why do I keep getting corrupted hard drives?

I'm wondering if changing to an LSI card would have any benefit... as I'm running out of ideas. The issue seems too low-level for the RAMdisk mod to be the cause... and back when I used to run it as a true main rig, it never did this (then again, that was on the 9500 controller, not the 9550).

[ratdude747]

Re: Ratdude's Main rig V2

She's been out of comission for several months... until last night (hopefully). Same random hard drive corruption issue. I replaced the following:

-RAID card (moved the BBU over as I didn't have a spare)
-SATA cables

No bueno, crash crash crash. But then I noticed something... the BBU was never staying charged. It was always "charging"... and the battery was verified good.

Bought a couple of used 9550's for cheap on fleabay, one arrived DOA (large scratches, a few SMD caps missing, caused a no POST when installed), the other seemed ok (both had dead batteries). Swapped the battery from my old one, and presto. Windows 7 installed without a hitch (no glitchy shit like before), and in 30 minutes of running the BBU finally reported "OK". My guess was the BBU controller was faulty, and causing issues with the write cache, which over time was corrupting the array more and more.

Time will tell if that fixed the issue. We shall see.

[ratdude747]

Re: Ratdude's Main rig V2

Finally did a proper full PSU recap using Nichicon HE's... and she's back and stable. Posting from her now.

FYI, there's two 470uf 16V caps, not 1

[ratdude747]

Re: Ratdude's Main rig V2

3300uf 6.3v 10mm x4 (Sanyo WG)
2200uf 16v 10mm x3 (Samxon GD)
470uf 16v 8mm x1 (Nichicon HM)

And the factory teapo mentioned above (220uf 50V 10mm)... which it appears I desoldered and resoldered, which would explain the cap sitting funny.

Maybe the PSU isn't the issue? Either way I probably should recap it again, this time using things that are better suited for PSUs?

I do still have the Ultra 600W from post #1... If I dare try that.

[ratdude747]

Re: Ratdude's Main rig V2

The ESR on Sanyo WG and Samxon GD might be a little too low for PSU use....but a questionable Teapo would be more suspect.

This was recapped with whatever I had in the stash years ago... maybe I need to redo it the right way?

[Topcat]

Re: Ratdude's Main rig V2

The ESR on Sanyo WG and Samxon GD might be a little too low for PSU use....but a questionable Teapo would be more suspect.

[ratdude747]

Re: Ratdude's Main rig V2

Pics of the cap in question:





Not sure if I knocked it loose last time I was in here, or if the bung bloated and caused this.

Also, I don't have any 50V 220's in stock, so no instant fixes (unless it is a 5VSB cap, in which case I do have 220's in lower voltages handy).

FYI, the recap I did last time was a mix of Sanyo WG and Samxon GD.

[ratdude747]

Re: Ratdude's Main rig V2

Spoke to Topcat, who suggested I look at the PSU. I checked, yup, I did recap it already, minus a 220uf 50V teapo between the primary and secondary heatsinks. It has come loose from the glue, so maybe that bloated on me and caused the issue?

Either way, I'll replace that cap at some point. Would that be the 5VSB cap? I've had 5VSB issues cause RAID controller headaches before (Main rig 3.5) so this could be the culprit?

[ratdude747]

Re: Ratdude's Main rig V2

No bueno. New card, redefined array, windows 7 install faults, upon reboot all 3 drives labeled "drive error". Delete them from the "device error" arrays, remake the array, and the bleeping revolving door spins again.

I will admit that I installed genuine DAZ loader right before the troubles began with the new drives... which caused a bootloader error that I tried to use bootrec.exe on the windows 7 dvd console to fix, which then caused the first failure of the array.

Perhaps something is wrong with my windows 7 driver? I'm running latest firmware and the latest driver (9.5.4 codeset); If Arch Linux supported 686 (PAE), then I'd just do that at this point and screw windows. Or perhaps I try windows 8.1 or something else?

Scratching my head here...

[ratdude747]

Re: Ratdude's Main rig V2

Stupid POS. New HDDs installed, ran great until it crashed the new win 7 install, only for the controller to mark all 4 drives as "drive error". Clearing that (redid array) causes the same windows 7 0x80300024 error.

Swapping controllers (I have a spare 9550sx), hopefully that helps.

[ratdude747]

Re: Ratdude's Main rig V2

Buying three of these:

http://www.ebay.com/itm/Western-Digi...ell-N963M-Tray

Got a decent deal on them...

[ratdude747]

Re: Ratdude's Main rig V2

She's croaked... windows 7 won't install without throwing an error.

Time to shop for drives...

[ratdude747]

Re: Ratdude's Main rig V2

Still comes up as "drive error", although I've found that its a flag on the RAID data that is only cleared by resetting the array (bye bye data). Drive is working.

Pulled the other two, not as tarnished but still pretty bad. Seeing if I can clear the flag without losing data... if not, reinstall time

[ratdude747]

Re: Ratdude's Main rig V2

Pulled the drive, PCB was badly tarnished. Cleaned it up and trying again...

[ratdude747]

Re: Ratdude's Main rig V2

If it weren't bad luck, you'd have no luck at all.

Sad but true. Ran great, went to another room for 20 minutes, find it on the "no operating system found" screen...

That's what I get for using bargain basement used drives in a RAID 0.

[Topcat]

Re: Ratdude's Main rig V2

If it weren't bad luck, you'd have no luck at all.

[ratdude747]

Re: Ratdude's Main rig V2

... And my raid 0 just died. one drive reports "drive error".

Damn Raid 0... now to find either another WD 160Gb (or whatever brand/model I used) or 3 of something else. Velociraptors?

[ratdude747]

Re: Ratdude's Main rig V2

Since teaming is dead, new challenge I'm considering: Windows 8.1 modded. Windows 8.1 is the last windows it will support (due to the lack of NX bit support, that didn't come until Xeon Irwindale), and since I've already broken every logical rule with this thing, it would be fitting.

Or not. IDK. 7 is a better desktop OS (and has "classic" games free without ads!).

[ratdude747]

Re: Ratdude's Main rig V2

Bummer:

https://www.intel.com/content/www/us...000005519.html

If I want teaming, I'd have to go back to Vista. Damn you PCI-X obsolescence. Hence why ever since updating drivers and rebooting, my NICs are dead.

At least V4.1 should work (IIRC it's dual NICs are PCIe and thus supported).