[SOLVED] Let's work this one out together - Noobs and Old Timers

[atlanticjim]

Hey everyone.
For the past two weeks, I have had a machine in the shop that I just could not figure out. When I finally got to the cause and solution, I wanted to share it here. I see that there are a few new techs active on the forum so I thought I would share my experience in a new way. Sort of a classroom experience where we demonstrate how we logic it out the solution rather than stabbing in the dark. Hopefully the noobs will see how we more mature techs approach a problem and learn from our mistakes and successes. I will present the problem from the start and would like to hear how you would approach it. In turn I will post the results of your tests. I know that there were missteps I made along the way and I will share that with you too. Sound interesting? Lets dive in.

The Chief Complaint
The 20 year old son of a good client calls me for help. His computer keeps shutting down after about 1 1/2 hours of operation. He says he sees a blue screen flash by, and then it reboots. After it reboots it works again for about the same period of time. His friend has tried a few things, used a few tools but hasnt been able to figure it out. Something about pulling out the RAM and trying it. He left a folder on the desktop with the tools that they used.

The Patient
I knew that this kid was a gamer but I wasn't prepared for what I saw.
First thing I notice is that this thing is HUGE, a white case that looks like a StarWars Storm Trooper. Actually it is a Stryker case from CM Storm. He and his buddy built it about a year ago and it has been running just fine until about a month ago. He says that the OS was from his buddy and probably not legal. Aside from the contents listed below, I has a sealed water cooling unit for the CPU.
Here is the Speccy report on the contents.

Summary
Operating System
Windows 7 Professional 64-bit (possibly pirated)
CPU
Intel Core i5 2500K @ 3.30GHz 30 °C
Sandy Bridge 32nm Technology
RAM
8.00GB Dual-Channel DDR3 @ 665MHz (9-9-9-24)
Motherboard
BIOSTAR Group H61ML (SOCKET 0) 33 °C
Graphics
ATI AMD Radeon HD 6800 Series (Gigabyte) 38 °C
Hard Drives
119GB M4-CT128M4SSD2 ATA Device (SSD)
(there is also a 500GB hard drive in the case but disconnected.
Audio
High Definition Audio Device​

Diagnostics
Let's hear how you would start....

 

[codegreen]

OK, I'm game. The first thing I would do is test the hard drive. First the SMART short test, and, if it passes, the extended test.

 

[markcuk11]

I would start with overheating issues or the SSD firmware bug from a while back!

 

[Cambridge PC Support]

Test the memory as part of standard diagnosis - and HDD as CG said.

Also, stop the machine from rebooting on Blue Screen and check out what it actually says.

 

[Kitten Kong]

I'm game also.

Re the blue screen, I would restart the machine with disable automatic restart.

Let the machine be used, and when the blue screen appears note the code, and then check what exactly the error code is.

This would aid me if its a hardware or software issue, ie buggy driver, memory diagnostics, or hard drive diagnostics.

Even using blue screen viewer, to check previous bsod's.

 

[ComputerRepairTech]

Even using blue screen viewer, to check previous bsod's.

This is the correct first step in my opinion. Not only this but you should verify how consistantly the system BSODs at 1 hour and 30 minutes by comparing bsod to system logs.

Edit: Oh right this is for newbies too, the reason why I would be doing this is to determine how I will proceed with my diagnosis. If the system BSODs every single time almost exactly at 1 hour and 30 minutes im going to determine whats happening at that time. First steps would be to check task scheduler and probably have process monitor running before that mark.

 

[Martyn]

All of the above but as you mentioned it I would like to see the tools he used from the desktop. Rebooter Pro isn't in there is it, or something like it?

 

[nightkingdoms]

Just from the description I'd guess it had something to deal with memory. I would think if it was HDD-related it would throw an I/O error not a halt.

I'd first ask them what programs they're using because what I'd guess isn't that the memory is bad (at first) but a program that updated and is addressing the memory incorrectly. Both would cause a halt (e.g. BSOD) but using it 1.5 hours into it and it happening, I'm thinking software then hardware.

 

[Bryce W]

Re the blue screen, I would restart the machine with disable automatic restart.
Even using blue screen viewer, to check previous bsod's.

I second this. Seems to be the first step since the machine is BSOD'ing

 

[SAG]

Windows 7 Professional 64-bit (possibly pirated)...
....
....
....
Let's hear how you would start....

By walking away.
If he got the OS from his friend it isn't possibly pirated, it is more than likely pirated, and I want nothing to do with it. At that point the issue could be with the loader/crack (or anything that tagged along with it) or something from one of the other pirated pieces of booty he has on there. I'd wash my hands of it before even digging in.
But that's me.

 

[nlinecomputers]

Me too. I am all over checking on what that BSOD actually is.

 

[NYJimbo]

By walking away.
If he got the OS from his friend it isn't possibly pirated, it is more than likely pirated, and I want nothing to do with it. At that point the issue could be with the loader/crack (or anything that tagged along with it) or something from one of the other pirated pieces of booty he has on there. I'd wash my hands of it before even digging in.
But that's me.

I was waiting for someone to say this. If I could confirm that its a pirated O/S I would tell them I cannot work on it. I've seen my share of machines with cracked/hacked O/S that after all the work you have done, you cannot install needed updates or something in the O/S isn't right and it stops you right there.

Sure, most of the previous posters have the right idea on actual diags, but for me I will not touch a system showing POSSIBLE O/S issues if the O/S is pirated.

 

[Metanis]

What about the System Event Log?

I would check the System Event Log first. If there is no record of the BSOD and reboot in the Event Log I would tend to look at the hardware. However if the machine is rebooting because the OS is pirated that will be a plainly readable event in the log.

 

[NYJimbo]

However if the machine is rebooting because the OS is pirated that will be a plainly readable event in the log.

Are you saying there would be a event in the log telling you that it rebooted because it is pirated ?

 

[seedubya]

While bearing in mind that this could be hardware related my initial thoughts are that if it's crashing every 90 minutes, pretty much like clockwork, then it's probably software related. So I would proceed as follows

Step 1: Disable Automatic Restart on System Failure - so that the next time it BSODs you can see the actual error
Step 2: Run Autoruns and check for anything unusual especially in the Scheduled Tasks tab.
Step 3: Try BlueScreenView and/or WhoCrashed to examine the crashdumps

Can't proceed any further without the results of the above

As for pirating - I would not refuse to work on such a machine. If the pirating is causing an issue then that can be resolved by getting the customer to purchase a new OS. If pirating is not the cause or is not causing any other problems then it's not my business other than to inform the customer of the risks of running pirated software and to advise them to get legal.

 

[NYJimbo]

As for pirating - I would not refuse to work on such a machine. If the pirating is causing an issue then that can be resolved by getting the customer to purchase a new OS. If pirating is not the cause or is not causing any other problems then it's not my business other than to inform the customer of the risks of running pirated software and to advise them to get legal.

You would have to convince them to do a TOTAL O/S reload on top of getting the legal O/S. But you also do not know if there is anything else wrong with the machine. So selling them a $100+ O/S and maybe a $80+ reload and then if anything else is wrong you have to deal with that too. I find that people who pirate O/S from their friends will RARELY decide to come clean and buy a new O/S from me, most will go back to their friends for advice because their friends machine likely works fine with the cracked/pirated O/S.

If I suspect an issue may be O/S related and the O/S is pirated I tell them I cannot work on the machine, offer them options and move on. Too many machines coming in without pirated O/S, I can afford to walk away from a few.

 

[SAG]

As for pirating - I would not refuse to work on such a machine. If the pirating is causing an issue then that can be resolved by getting the customer to purchase a new OS. If pirating is not the cause or is not causing any other problems then it's not my business other than to inform the customer of the risks of running pirated software and to advise them to get legal.

And that's cool too.
Can be a good way to make a few bucks.

To the OP, my first check would be the memory or ssd.

 

[Larry Sabo]

I hate gaming systems, because they are usually tweaked within a nanosecond of crashing, and I don't have known-good parts to swap in (high-wattage PSUs, required high-speed RAM, gaming video cards, etc.). If forced to work on it, or because I hate to be beaten, reviewing the BOSD history with Blue Screen View would come first. Right away, I am suspicious of hardware as the root cause, so would probably run a live Linux CD to rule out the hardware before chasing phantom software/system symptoms. If Linux runs without drama, I'd remove/disconnect unnecessary components to get it down to a bare-bones system that runs reliably. If it still BSODs, I'd start swapping in known-good components until it runs reliably. If that comes to a dead-end, I'd condemn the MB, curse myself for having wasted so much time on the $#^% thing and ask the owner to remove it from my sight.

 

[YeOldeStonecat]

Ugh....a home built cloner with some budget parts...Biostar motherboard...double UGH. Who knows what compatibility issues there might be, if some kids built it....I have doubts of them checking for issues between hardware. Just grabbed a buncha parts and slapped them together.

So yeah gotta get a hold of what those blue screens are. It's only logical...you want to see what exact blue screen there is. Especially if they're the same each time...it practically points a flashlight on the issue for you!

RAM...did they get RAM specifically approved compatible for that...ick..I hate saying it..."BIO-STAR" motherboard? Or did they just grab any ol RAM that they though was the right speed? Does the RAM have any higher voltage requirements?

How about the power supply?

You can check the HDD...but it's a Crucial M4...I'd doubting it. Might be a compatibility issue between it and that...ugh....BIO-STAR motherboard though.

I'm doubting there's an exact 90 minutes between the reboots...I think that's just an estimate of a broad time period, so I'm not focusing on that.

Overheating CPU...possibly poorly setup heat sink.

 

[Lewis]

Process of elimination

While bearing in mind that this could be hardware related my initial thoughts are that if it's crashing every 90 minutes, pretty much like clockwork, then it's probably software related. So I would proceed as follows

Step 1: Disable Automatic Restart on System Failure - so that the next time it BSODs you can see the actual error
Step 2: Run Autoruns and check for anything unusual especially in the Scheduled Tasks tab.
Step 3: Try BlueScreenView and/or WhoCrashed to examine the crashdumps

Can't proceed any further without the results of the above

As for pirating - I would not refuse to work on such a machine. If the pirating is causing an issue then that can be resolved by getting the customer to purchase a new OS. If pirating is not the cause or is not causing any other problems then it's not my business other than to inform the customer of the risks of running pirated software and to advise them to get legal.

My steps would resemble these to the T .....as for the pirating ...I as well would educate my client instead of refusing the work.