Hi Duke, I think, Ignite should be cope with any issue. For example: looped thread which does not handle flag of interrupted.
But you always can to try do that, by way implement your segmentation behavior. Something like this: ignite.events().localListen(new IgnitePredicate<Event>() { @Override public boolean apply(Event event) { (new Thread() { Ignition.inite().close(); Ignition.start(new IgniteConfiguration()); }).start() return true; } }, EventType.EVT_NODE_SEGMENTED); Do not forget create new configuration, because Ignite changed state in it. On Fri, Nov 11, 2016 at 4:51 AM, Duke Dai <duke.dai....@gmail.com> wrote: > Hi vdpyatkov, > > Finally, I figured out not only one but all nodes become segmentation due > to > unknown VMWare infrastructure. > I changed failuredetectiontimeout/sockettimeout/networktimeout, and the > cluster survived last day, need more time to observe their behavior. > > I'm thinking SegmentationPolicy. > Why SegmentationPolicy.RESTART_JVM was provided(must work with > CommandLineStartup)? Any limitation/implication that soft-restart(in same > JVM) won't work? > > > Thanks, > Duke > > > > -- > View this message in context: http://apache-ignite-users. > 70518.x6.nabble.com/How-many-nodes-within-your-ignite- > cluster-tp8808p8892.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com. >