PDA

View Full Version : Back from the brink



TomNanos
09-28-2007, 09:23 AM
Well, I've been battling with some computer issues on my home machine for the past week or so. Everything from crashing apps for no reason at all to random reboots (again, for no good reason), and even corruption of about 50 RAW images. I got a chance to dig into it this week as things were getting worse (and the family pig roast we hosted this past Saturday was over...), and it seemed that 10% of the RAW images copied to the machine were being corrupted.

First up I thought it might be a filesystem issue. So I did full chkdsk scans of each drive on the system - both internal and external. No issues at all. Hmmm....

After some digging, research and investigation on the computer itself, I first determined that my power supply was woefully underpowered for the hardware attached to it. It seems that a 250 watt power supply just wouldn't keep up with 3 internal hard drives, an internal DVD burner, 10 USB devices, 3 firewire devices and of course the motherboard/CPU/memory. Some quick calcuations on the power consumption came up with about 500 watts of power need! So that would explain some of the random reboots, and possibly some of the corruption.

So I ran down to CompUSA in Manchester (about 30 minute drive from home), got a 500 watt ATX supply (on sale for $60!), and got back home in time for the new episode of House. Installed it during the commercials, and all seemed right with the world. Until Firefox crashed. Damn, what else could be causing that? I've got plenty of power now. Maybe there's some corrupt DLLs in the mix too. Too late to keep looking, so I'll dig some more the next day.

Well, while at work I was talking with my boss about my issue. We were batting some ideas around, and he mentioned that maybe the low power condition affected one or both of the DIMMs. Interesting - that would fit with what's going on since it appears to be random, and the type of corruption of the RAWs usually indicates that one or two bits were flipped in the file itself.

So that evening I downloaded a pretty neat tool, Memtest86, which is a bootable CD that does a battery of tests on each address of RAM. So I burned the CD and rebooted. The tester began to do its thing, with tons of errors being reported. Hmmm, that may be it! So I shut it down again, pulled one stick of RAM out and ran the test again. More failures at the same addresses. Looked like the last two bytes of each address were being mangled. OK, next I checked the other stick. That one passed with flying colors! Yep, it looks like I had one bad 512MB stick of RAM. Once the tests finished up, I rebooted and all seemed right with the world. Did some test copies of RAWs and everything seemed to copy OK. So I restored most of the corrupted RAWs (I figure I lost about 50 in all that were also corrupted on the backups) and called it a day.

Only cost me about $120 to fix ($60 for the PSU and $60 for a new 1GB stick of RAM), and a few days of pulling my hair out troubleshooting. I'll probably be doing some kind of MD5 digests in the future to verify the data integrity and hopefully catch this before it happens again...

Live & learn...

Orgnoi1
09-28-2007, 09:31 AM
That COMPLETELY sucks... but at least you got it set up where its working... and got a couple upgrades in the interim...

TomNanos
09-28-2007, 09:39 AM
Yeah, it did suck, but at least I caught it. I'm kicking myself for not catching it earlier - hindsight is 20/20. I remember thinking it was odd that my portable USB drive wouldn't power up unless it was connected to a powered hub. Well, the underpowered PSU would explain that completely! Oh, well...

And yeah, I'm running a gig and a half of RAM right now...that should keep CS2 happy for a bit. :D

KDN
09-28-2007, 10:32 AM
Wow I actually read that like I understood any of it LOL I am NOT computer smart :) Sorry to hear you had all sorts of trouble, and I give it to you cause I probably would have been so mad my computer wouldn't exist any more :D Glad you got it all sorted out, at least I know who to call when I'm ready to throw my computer out the window LOL

JayVig
09-28-2007, 10:40 AM
Tom, You had a gig of ram before? How did lightroom like only a gig?

I'm on a 1.8gig cpu and a gig of ram and lightroom sucks the life out of my machine.

convergent
09-28-2007, 10:51 AM
Don't you wonder sometimes what people do that aren't geeks? I think of the number of hours that I spend working to keep the computers in my house operational, and the trouble I have doing it. Then I think of my other family members who know nothing about computers. How do they do it?

Orgnoi1
09-28-2007, 11:00 AM
Then I think of my other family members who know nothing about computers. How do they do it?

LOL they ask YOU... like they ask ME...

TomNanos
09-28-2007, 11:01 AM
Kadian - trust me, there's some carcasses in the basement of machines that, well, didn't quite make it. :D Steel toed boots do wonders against motherboards.... ;) :D

Jay - Lightroom ran decent on a gig of ram. CPU is a 2.66GHz P4. I had to turn off automatic writing of the settings to XMP to make it work fine, otherwise it would just bring the system to it's knees every 40 minutes or so. Looks like they fixed that in 1.2, which I have yet do d/l on my home machine. I've got 1.2 on my laptop at work (shhhhh...it pays to be an admin :D), and it works great now with auto-XMP on. Only have a gig on the laptop with a 1.3GHz mobile P4 and it works just fine.

Mike - I'll tell you what my family members do - they call me! It gets tiring, but my father finally installed PCAnywhere on his PC (my parents live in Florida), so I can remotely troubleshoot, even through the firewall here at work. But my sister calls me at least 3 times a week with questions or problems...

TomNanos
09-28-2007, 11:06 AM
Oh, yeah, and Jay, I've also got about 6GB of swap file space spread across the 3 of the 5 internal hard drives (I forgot I've got 2 SATA drives in the box too...so my assessment above was off a bit, but they were included in the power consumption estimate). Also, turn on the /3GB switch in the boot.ini file (switches the split between user and OS available memory from 2GB each to 3GB for user and 1GB for OS). That helped immensely.

JayVig
09-28-2007, 12:08 PM
Thanks for the lightroom help. I think that'll make a difference. I'm still not sure of all it can do because I haven't even used it. I did, however, get it to open once without my PC going tango uniform (military types will get that reference).

As for the family help... my father relied heavily on me when I was in the area, once I left for the military he was at a loss. One visit I sat down unbeknownst to him and dropped a VNC server on his machine, made his ip static, did some port forwarding arrangements on his router and registered a nice, free dynamic DNS. the next time he called I told him to standby just a minute and before he knew it I was in his PC. Of course he went berserk when he PC became possessed until I told him it was me - but that was half the fun.

Once I deployed though and I wasn't reachable - he just bought a Dell. Problem solved. Warrantied up the yin yang with 24/7 phone support and on site stuff. He's covered. I think the warranties cost more than the machine. haha.

zacker
09-28-2007, 12:17 PM
Don't you wonder sometimes what people do that aren't geeks? I think of the number of hours that I spend working to keep the computers in my house operational, and the trouble I have doing it. Then I think of my other family members who know nothing about computers. How do they do it?


easy, we dont go messin with our computers innards, installing stuff, uninstalling stuff... so nothing breaks.. and if it does and its too expensive to fix, we go out and buy a whole new one... it helps keep the computer industry alive.. lol Computers now are like the cardboard "Box" cameras, use untill they cant be used no more then go get another one.

zacker
09-28-2007, 12:21 PM
Once I deployed though and I wasn't reachable - he just bought a Dell. Problem solved. Warrantied up the yin yang with 24/7 phone support and on site stuff. He's covered. I think the warranties cost more than the machine. haha.



see? thats what we do! :D

JayVig
09-28-2007, 12:24 PM
Well Craig it was probably because I was going to bill him by the minute for phone support. Nothing like sitting on the phone with pop for 45 minutes on a friday night while trying to ignore the scowl of my better half wondering when we're headed out to dinner. So I figured I should profit off the calls. It's gonna be my inheritance anyway so I thought it would be easier for him to give it to me in little bits instead of a lump sum






None of the above is even remotely true. Especially the part about receiving "that look" from my lady. She's sweet and would never do such a thing.





Part 2 is also untrue but you never know when they'll find a way to read this...