Den 11 maj 2014 08:09 skrev "Hervé BOUTEMY" <[email protected]>: > > Le mercredi 7 mai 2014 11:35:29 Olivier Lamy a écrit : > > Hi Guys, > > I did some upgrade today. > > Align jdk for ubuntu nodes to be the same as the master. > > > > Install the Jenkins monitoring plugin: https://builds.apache.org/monitoring? > FYI, Jenkins admins are the only one who have access to the result > Not a problem per se: it's good to know that you guys now have some recording > on key JVM information (and perhaps sometime sys admins will like JVM based > service administration once they'll use such tools :) ) > > > > > Andrew will run some scripts to cleanup some configuration. > > I found previously some jobs with a scm pooling defined to: * * * * * > > (this means using some threads all the times to check scm changes) > > Furthermore, some jobs doesn't have any discard previous builds > > configured this tend to use a lot of disk space for nothing!! > +1 I'm sure some of the actual problems are caused by improper jobs > configuration by people not aware of the consequences > > BTW, will the script fix problems without reporting to job owners? or will job > owners be warned that they did something wrong?
Another tool to find bad job configurations is the Configuration Slicing Plugin. I don't know if it's installed. It doesn't cover all configurations, but the key ones are there. It shows one configuration parameter at a time for all jobs, grouping jobs with the same configuration together. This makes it easy to find jobs that stand out in some way. It also let you change one configuration option for multiple jobs at once. > > It's good to see such work happen and explanation on the ml: I hope we'll > rapidly see the improvements on the stability of the service > > Regards, > > Hervé > > > > > > > Cheers >
