Sorry I was not clear earlier. mesos-slave is run as docker container and containerizer is also docker. I saw this in another thread, https://issues.apache.org/jira/browse/MESOS-2115
Was the link related to the problem I described? On Fri, May 29, 2015 at 4:56 PM, Bjoern Metzdorf <[email protected]> wrote: > Siva, > > do you need to enable checkpointing? > > http://mesos.apache.org/documentation/latest/slave-recovery/ > > Regards, > Bjoern > > On May 29, 2015, at 8:01 PM, Sivaram Kannan <[email protected]> wrote: > > > If the slave recovers within the timeout, why should the the tasks on the > node be killed? > > On Fri, May 29, 2015 at 1:15 PM, Alex Rukletsov <[email protected]> > wrote: > >> Siva, >> >> yes, this is intended behaviour: keep tasks running and give the Mesos >> Worker some time to re-register. You can adjust this timeout via >> --slave_reregister_timeout, but keep in mind 10 min is the minimum. >> >> On Fri, May 29, 2015 at 8:04 AM, Sivaram Kannan <[email protected]> >> wrote: >> >>> >>> Hi , >>> >>> If I restart a mesos-slave service in a node, does it brings down the >>> apps running in those nodes. The apps are coming up again after mesos-slave >>> service is up, but wanted to confirm whether that is the expected behaviour. >>> >>> Thanks, >>> ./Siva. >>> >> >> > > > -- > ever tried. ever failed. no matter. > try again. fail again. fail better. > -- Samuel Beckett > > -- ever tried. ever failed. no matter. try again. fail again. fail better. -- Samuel Beckett
