Thanks! On Mon, Jan 18, 2016 at 12:23 AM, Marco Massenzio <[email protected]> wrote:
> Hi Michal, > > 1. While watching some talk I've heard that maybe in the future ZooKeeper >> won't be needed. Is this still planned? >> > > At some point there has been talk of moving towards using etcd instead of > ZooKeeper: y > ou can look into Jira[0] > , and it seems that MESOS-1806[1] is the one that has received the most > attention/activity > . > > Others may be able to provide more detailed guidance, but the impression I > have is that it may be some time before this becomes available as a > Production-ready alternative. > > > >> 2. We're using mainly quite large boxes (>= 20 CPUs, >= 48GB RAM). Is it >> advised to put Mesos master and warm backup nodes inside ZooKeeper's >> cluster? (just to avoid wasting resources). >> >> There is really no reason not to have Master/ZooKeeper servers > co-located - in fact, this is the way DCOS CE is deployed in AWS (or, at > least, used to last time I looked into it). > > Hope this helps! > > [0] > https://issues.apache.org/jira/issues/?jql=project%20%3D%20Mesos%20and%20text%20~%20%22etcd%22 > [1] https://issues.apache.org/jira/browse/MESOS-1806 > > > > -- > *Marco Massenzio* > http://codetrips.com > > > > -- BR, Michał Łowicki
