Hi Andrea, would it make sense to add `-XX:+ExitOnOutOfMemoryError` to surefire/failsafe JVM arguments? The idea is that we would not get good results from the tests anyhow, so we might just as well fail completely rather than continue.
zoran On Thu, Jan 23, 2020 at 2:33 PM Andrea Cosentino <[email protected]> wrote: > > We cannot cleanup machine that are going out of memory. The builds were > failing in the last days for this reason. > > Il giorno gio 23 gen 2020 alle ore 14:31 Zoran Regvart <[email protected]> > ha scritto: > > > Hi Cameleers, > > there's a discussion on the builds list[1] and JIRA issue that > > mentions Camel[2]. I wasn't aware that our builds would fail and still > > consume resources on the build nodes. > > > > I guess it's hard unless we can look at the processes on the node > > itself. Did anyone observe this in any other way? > > > > If this is the case, does anyone have an idea on how to cleanup after > > the job is done (without disrupting other jobs)? > > > > zoran > > > > [1] > > https://mail-archives.apache.org/mod_mbox/www-builds/202001.mbox/%3CED76C2EF-1AED-4046-8E6A-427E7B3C1A64%40apache.org%3E > > [2] https://issues.apache.org/jira/browse/INFRA-19751 > > -- > > Zoran Regvart > > -- Zoran Regvart
