Get the whole picture

When you have as many computers to maintain as I do, you try to automate as much as is practical.  I do as much remote maintenance as possible from my office so as not to waste time schlepping all over campus.  (It’s only PART laziness, I assure you!)  However, sometimes there is no substitute for on-site inspection of a problem.

This one particular computer in the Levitt Center has been brought to my office three times because "it’s not working".  I get various reports, from "it’s not working" to "people can’t log in".  Nothing terribly specific.  Each time it comes to my office, I re-image it and send it back.  A few days later, I get another report.

This time, I returned it to Levitt myself.  I hook it up.  It fails to detect network connectivity.  I unplug/re-plug its cable.  I know there’s nothing wrong with the computer, because it was working FINE in my office a scant 5 minutes ago!  I go downstairs and inspect the switch.  There’s a link light, but it only shows a 100Mbit connection.  (This computer is capable of 1000Mbit, as is the switch.)  I go back to my office and check the switch configuration.  The port is properly configured. 

So I take another cable with me and go back to Levitt.  I plug in the computer using the "new" cable and it works.  Success!  But WHY?!  The cable to which it was previously connected was "store-bought", not "hand-made".  Almost none of these things fail!  So I pulled it out of the cabinetry and, on a lark, I look at the end before I toss it in the trash. 

Lo and behold, there is a single, tiny grain of sand lodged in one of the 8 pin slots.  I removed the grain, re-attached the cable and it works just fine now.  No amount of remote management–or even instruction of anyone else in the troubleshooting of this problem–would have led me to this result.  Sometimes, you just have to go see for yourself.