Tuesday, December 5, 2017

Take every bug report seriously (Part 2)

Introduction


Many engineers value rationality above reality. Even if they see something is happening in front of their eyes, they will want to discard it as a "one-off" case if they can't find a reason behind this. 

I have seen this during several instances. Many a times, bugs filed by QA Engineers are discarded by Dev Engineers as "Not Reproducible" if they cannot find a reason why the bug would have happened in the first place; or if they cannot reproduce it on their systems ("It works on my machine!").

In this series of blog posts, I will post many stories that helps us (the engineers) understand the importance of treating every defect seriously.

Story #1


Story #2

Eight years ago I was working as an escalation tech for a high end computer manufacturer that is no longer in business. They specialized in very expensive home rigs. One of my techs received a call from a user on the east coast complaining that their computer would reboot for no good reason every evening around the same time, give or take a couple hours. The tech checked all the obvious settings (BIOS, Power Management, etc.) but could find nothing that was telling the computer to reboot at any particular time. It was running <shudder> Win95, not NT so there were no OS timeouts available. 

They then started working the hardware angle -> remove all add-ons and see if it still has the same problem. Disconnected the printer, scanner and joystick and ended the call. A couple hours later user calls back saying the computer had just rebooted on its own again. The new tech followed up with the first tech to see what else needed doing and they ended up escalating to the supervisor for an RMA of the power supply or mobo. Sup decided to RMA both because he could not prove which one was bad.

Customer got new parts, installed them himself and was up and running for the entire weekend. Monday rolls around and he calls back in the late evening (7 or 8PM) with the same problem. This call got escalated straight to the sup who decided we needed to RMA the entire tower and have the repair team take it apart. A week later customer gets his system back with all new components inside. The only thing that was not replaced was the case (serial number). They even replaced the front bezel of the case in case it was a malfunctioning reset switch.

That night he calls in again WITH THE SAME PROBLEM. At this point, my sup decides there is no way the problem exists within the confines of that computer case so he hands it to me to try and figure out (he didn't like me very much). While I'm talking to the customer (about 50 minutes into the call) I hear a noise in the background. He immediately lets me know that the computer has just rebooted. The noise that I heard did not sound anything like what you would expect to hear from a computer that just came out of the box, so I asked him what was the noise that I had just heard. He informs me that his wife had just come home from work and was "finishing her business in the bathroom".

I pondered this new bit of information for a while. I asked if she works a regular 8-5 type job and always comes home at the same time, every weeknight. His response was that sometimes she works late and then gets to come home early the next day, but that it never varies by a couple hours.

After the computer was back up in Windows I asked him to go flush the toilet and then come back and check the computer. He did and sure enough, it was in the process of restarting. I then asked him to check for any wires leading into the tank of the toilet. He informs me that there is no tank on his toilet. He "only buys the best" so instead of having a tank, he has a water pump that pumps water into the toilet from a water reservoir in the basement that is kept at room temp (also supplies water for the bathroom sink and indoor watering system for all their potted plants - more money than he knows what to do with, obviously).

The water pump just happened to be on the same circuit as the outlet that his computer was plugged into. Despite the use of a surge protector, the water pump kicking on put enough of a drain on the circuit to cause the computer's power supply to momentarily die (this was 8 years ago, before digital power supplies), but was not a surge so the surge protector never tripped. I had him move his computer desk to another wall in the same room and plug into one of it's outlets instead. This moved him to a different circuit and solved the problem.




[NOTE: This popular story has been taken from Internet. I do not know the very original author, so have not attributed to him/her.]

No comments: