Mail Server

It all started on a Monday … This reseller told me that there was a customer of his that was getting mail server errors … So I tried sending from/to accounts I have on that server … It worked. I also sent messages to the guy that was reporting the problem … They all worked. But, he tried sending, and he couldn’t do it, so then I thought that maybe I had some sort of privileges since I was on the same network … Regardless, then the phone rang … Someone trying through webmail wasn’t receiving messages … Another IM popped upped … Damn! Something was a amiss … I immediately contacted the mail server guy and he found that the virus scanner was crapping out on a corrupted database … He then thought it best to upgrade the virus software to fix the problem … All was fine and good … until the next day … Now the thing was core dumping … Various other errors, and the machine would quit under load and boot … The thought was that Perl should be upgraded due to the new virus software … While that was compiling, it was crapping out, so then things got real fun … So he figured that we should just upgrade the whole box and bring everything current … but the machine seemed to behave respectably if we left it alone … We did get a new kernel built, but trying to right everything else was resulting in Bus errors and more and more core dumps … At this point, the problem seemed to point to memory although the memory in this machine has been working for the last year … We decided to leave the machine running and stop the IMs and phone calls until I could test the memory in the wee hours … I did that yesterday morning and there were multiple errors so lo and behold, yes! It was the memory … So having that tested, I had to go another day running the bad memory until I could purchase new memory and install it. So, on my way to work I stopped at CompUSA and picked up 1G of memory … When I got home, I first tested the new memory in another machine … I checked it for over an hour and it passed all tests … Then I installed it into the mail server and have been up ever since … Then I tested the RAM again in the test machine … The stick that went bad wasn’t even the one I thought would have went bad … The 256 stick was bad, and the 512 is good … So, now we need to fix up the rest of the mail server and all will be well … I can’t tell you how I have felt the last week with all 3 of these scenarios hitting at once … At one point I almost gave up … but, that’s not like me …

Leave a Reply

Your email address will not be published. Required fields are marked *