Sorry, previously did send from @apache.org. Mesos is exhausting Jenkins resources: https://issues.apache.org/jira/browse/INFRA-16030 You execute Maven 23 runs. You had to already notice that your build takes longer, it allocates 4 or 5 runs in one machine but the problem is that each CI machine has 2 executors.
On Wed, Feb 14, 2018 at 1:33 AM, Tibor Digana <tibor.dig...@googlemail.com> wrote: > Mesos is exhausting Jenkins resources: > https://issues.apache.org/jira/browse/INFRA-16030 > You execute Maven 23 runs. You had to already notice that your build takes > longer, it allocates 4 or 5 runs in one machine but the problem is that > each CI machine has 2 executors. > > On Tue, Feb 13, 2018 at 11:17 PM, Olivier Lamy <ol...@apache.org> wrote: > >> Hi >> A bit complicated with timezone.... >> Let's discuss here in async mode :-) >> Cheers >> Olivier >> >> On 13 February 2018 at 10:04, Tibor Digana <tibor.dig...@googlemail.com> >> wrote: >> >> > Can we discuss it in the evening? >> > I am in the office now. >> > I will explain, but I normally don't have much time. >> > >> > On Mon, Feb 12, 2018 at 12:32 PM, Olivier Lamy <ol...@apache.org> >> wrote: >> > >> > > I don't know if it's a good idea to try running performance test on >> asf >> > > nodes... >> > > they are used by every project so you never know how it is used and >> for >> > > what, what is the configuration (virtualised or not can make a huge >> > > difference) >> > > >> > > On 11 February 2018 at 18:58, Tibor Digana <tibordig...@apache.org> >> > wrote: >> > > >> > > > I really have a problem to run my 11 parallel jobs on our Jenkins >> > because >> > > > Messos runs on most of the Ubuntu machines. Each Messos build takes >> > cca 1 >> > > > hour to complete. >> > > > This does not make any problem for you? >> > > > The Messos is triggered by a scheduler and not by a commit. >> > > > >> > > > Now I have to decide if I should have no limit or use only ubuntu-1. >> > > > The machine ubuntu-1 seems to be the most powerful machine and fits >> to >> > my >> > > > needs. >> > > > The problem is that 11 parallel jobs is quite a lot anf then it >> happens >> > > > that all jobs run on the ubuntu-1, the machine is overloaded and >> > > > performance test fail. >> > > > >> > > > So what should I do? >> > > > The IRC of INFRA is not available and I do not have free machines >> for >> > my >> > > 11 >> > > > jobs to run for one hour. >> > > > >> > > > Thx >> > > > Tibor >> > > > >> > > >> > > >> > > >> > > -- >> > > Olivier Lamy >> > > http://twitter.com/olamy | http://linkedin.com/in/olamy >> > > >> > >> > >> > >> > -- >> > Cheers >> > Tibor >> > >> >> >> >> -- >> Olivier Lamy >> http://twitter.com/olamy | http://linkedin.com/in/olamy >> > > > > -- > Cheers > Tibor >
