note that if I reboot all the automation machines (Unix master, Windows slave, Windows agent on which automation tests run compiles and text executions are very fast. I have six to twelve Jenkins jobs running every night.
However, over a period of a few days, performance will bog down and slow up to the point that the longest running job (optimally at 8 hours) will take up to 20 hours or so!
So, I reboot again and things speed up for a while and then bog down.
Last Friday, I noted the hang ups again and this time I noticed on the Jenkins dashboard that it is during compiles there is a hideous slowdown. I don't know this is always true and I will look at prior builds today to see if there is a pattern.
Before running the tests, there are jobs that compile the code and clean out "gunk" from output directories.
Any general guidance as to debugging this? I lack the experience and knowledge still to be able to organize and target my debugging efforts as well as I should.
I don't think I should have to depend on random reboots to maximize Jenkins performance.
Oh, here's the other big problem and this may a very large part of the problem. I 'm working in a changing environment where I have been limited in what I can do to update software as I should because of upper level decisions. I should be running the latest stable build of Jenkins and all software but I am not. That should be my first effort in solving the performance issue, right? I am just going to try to update all software myself at this point.
Thanks for listening.