Please provide full thread dumps from all the nodes at the time it happens again. In particular I’m interested in what is going with the thread named "ipc-shmem-gc”. In addition to the thread dumps attache the logs from all the nodes for investigation.
As a workaround you can disable shmem as it was suggested before. Finally I would recommend upgrading to the latest Ignite 1.6.0 version. — Denis > On Jun 24, 2016, at 3:30 PM, bintisepaha <[email protected]> wrote: > > JDK 1.7, 16 node cluster, each JVM 10 GB heap and 4 GB off heap. 4 nodes run > on 4 linux boxes, each box has 64 GB memory. the heap is under-utilized when > this happens. so doubt GC will cause it. > multiple clients connect to it. clients cannot exit properly either with > System.exit() or Ignition.stop(). > > We did not have the issue on our end until this week, but then something > could have changed with ports in our environment. We cannot reproduce > ourselves in UAT it only happens for us in Production. > > > > > > -- > View this message in context: > http://apache-ignite-users.70518.x6.nabble.com/System-exit-not-exiting-cleanly-locked-on-IgnitionEx-IgniteNamedInstance-tp5814p5876.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com.
