torsdag 6 februari 2014

When troubleshooting fails and how to solve it

Troubleshooting computers starts with running a diagnostics test to identify the faulty component in the computer. Because it must be a component that is broken right? I ran a test with 20 computers who needed to be checked. Half of them showed what was wrong in the diagnostics and parts could hence be replaced to get them to work again. The other half just shut themselves down after a long time. We didn't have time to stand and wait for error messages prior to the shutdown so we didn't knew if there where any. Since the diagnostics was doing extensive tests on the memory and these tests took most of the total time for the diagnostic we thought it could be the memory's that made the computer shut down. We therefore tried with memory's we where sure worked but got the same result. Since the memory test was the last one in the diagnostics and everything else had passed we knew all parts of the computer had passed the test.

So basically everything were fine with the computers, except they shutdown without any reason.  So if nothing was wrong with the computers why did they behave like this?

The solution:
We now knew that we were dealing with a common problem and that there was nothing wrong with the hardware, or at least it didn't seem so. Something made the computers shutdown. What could it be? In what situations do computers shutdown? Well, it could be something with their the power supply but since they all had working battery's and where all connected to working AC adapters this was highly unlikely.

Maybe it was their processors? When it gets overheated it shuts down and hence so does also the computer. But why would it get overheated? All computers had a lot of space around them and hence no problems with the airflow externally. But what about internally? Could it be dust in the processor fan? That would explain the overheating. We borrowed a compressor from another part of the school and blew them all hard in the out and intakes of the fans. I wish would have had some sort of mask on me because three years of collected dust did not do wonders on my lungs. Since clearing the fans seemed like something that was needed I didn't stop at those 10 I was troubleshooting. Instead I kept going until all 50 computers in the room were cleared from dust. Also, I should have used ear protection because that compressor made a lot of noise.

When finished, all 10 passed the tests and I could restore windows 7 images on them and put them on the "ready to be served" shelf.

Dust really can cause problems. Blowing it out can really make a difference. Noise from jammed fans, lowered lifetime of components due to high temperatures and sudden shutdowns can all be avoided by blowing with the compressor for a while.

Sometimes you have to think a bit out of the box to solve problems. It's not always obvious what the problem might be, especially if you haven't thought about it before or have any previous experience. Trying to find a problem from a different angle can really pay off.

Background:
I'm practicing at the IT-department at a high school called Österlengymnasiet in southeastern Scania, Sweden. The job is more of a solving problems one than just fixing computers. The main goal, which is called "one to one" is to make sure all students got a working computer.

Students get's a computer when they start their education and these are expected to last until they graduate. This is of course not the case, stuff breaks. Depending on the model, different components more often breaks and hence faults in the design are shown.


Inga kommentarer:

Skicka en kommentar