+1, when server hits an issue it cannot recover and continue to work meaningfully, it is best to shut down or restart. That approximate crash failure behaviour ( which is a good thing).
On Tue, May 10, 2016 at 11:40 AM, Ramith Jayasinghe <ram...@wso2.com> wrote: > HI Azeez/Sameera, > > Having encountered some OOM issues (relating to HZ) I looked at what > Hazelcast does in face of GC errors [1]. It will drop all connections, > threads, and shuts it self down. > > Now, Think about one of our server's ( say MB) face a OOM and embedded HZ > instance shuts it self down. I propose when this happens entire server > should stop its processing intelligently ( if not shutting down). > Rational is one of the sub-system in the server shutdown while other parts > of the server don't care, which is wrong. As for MB, it tries to stop > processing in a meaningful manner and I argue this should be something any > server should do if they use HZ clustering. > > thoughts? > > > [1] > https://github.com/hazelcast/hazelcast/blob/master/hazelcast/src/main/java/com/hazelcast/instance/DefaultOutOfMemoryHandler.java > > -- > Ramith Jayasinghe > Technical Lead > WSO2 Inc., http://wso2.com > lean.enterprise.middleware > > E: ram...@wso2.com > P: +94 772534930 > > _______________________________________________ > Dev mailing list > Dev@wso2.org > http://wso2.org/cgi-bin/mailman/listinfo/dev > > -- ============================ Srinath Perera, Ph.D. http://people.apache.org/~hemapani/ http://srinathsview.blogspot.com/
_______________________________________________ Dev mailing list Dev@wso2.org http://wso2.org/cgi-bin/mailman/listinfo/dev
