We have been running 0.14.0 on production clusters at Twitter and have found a small number of bugs. As such, I will be sending out a VOTE on a 0.14.1 release that will be based off of the 0.14.1-rc1 tag.
The most important of these bugs are: MESOS-662: When running Mesos with cgroups memory enforcement enabled, it is possible for an OOM event to cause a deadlock situation in the Linux kernel. MESOS-420: This has crashed the master process on rare occasions only in our large production clusters and as such it was missed during our testing in a staging cluster. MESOS-711: It's possible for the master to "over-subscribe" slaves (offer more resources than the slave contains). This manifests primarily when upgrading from 0.13.x to 0.14.0 on a running cluster. We've also included some reliability improvements to slave recovery related to edge cases (we discovered these by killing slaves at frequent intervals to stress test recovery). These are MESOS-679 and MESOS-680. The anticipated release JIRA changelog is here: https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12311242&version=12324994 Ben On Tue, Oct 15, 2013 at 6:46 PM, Vinod Kone <[email protected]> wrote: > Hi, > > I'm happy to announce the passing of 0.14.0 vote with 3 +1 binding votes, > no -1 and no 0 votes: > > +1: > Benjamin Hindman (binding) > Benjamin Mahler (binding) > Dave Lester (binding) > > -1: > > 0: > > Please find the release at http://www.apache.org/dist/mesos/0.14.0 (or > preferably please use a mirror from http://www.apache.org/dyn/closer.cgi). > > We've also uploaded the mesos-0.14.0.jar built for Java 6 that should be > accessible via Maven Central as well. > > Thanks everyone! > Vinod >
