Hi Yitzhak,

You can configure specific timeout in discovery SPI or increase common
timeout IgniteConfiguration#setFailureDetectionTimeout.
But long timeout lead to stop grid on time the timeout

If you want to handle segmentation event:

ignite.events().localListen(new IgnitePredicate<Event>() {
    @Override public boolean apply(Event event) {
        System.out.println("Execute custom logic...");

        return true;

But what will you do?
You cannot wait until something there, because cluster already segmented
(data will be rebalanced on other nodes) the node.

On Thu, Oct 13, 2016 at 12:40 PM, Yitzhak Molko <yitzhak.mo...@symbolab.com>

> While I didn't configure any network segmentation properties
> (SegmentationResolvers, SegmentationResolveAttempts,
> SegmentCheckFrequency etc.) node is been shutdown from time to time:
> WARN : [discovery.tcp.TcpDiscoverySpi] Date=2016/10/13/07/42/52/009|Node
> is out of topology (probably, due to short-time network problems).
> WARN : [managers.discovery.GridDiscoveryManager]
> Date=2016/10/13/07/42/52/009|Local node SEGMENTED: TcpDiscoveryNode
> [id=4b3349f5-fda0-4e9d-a528-c8b5f4401717, addrs=[0:0:0:0:0:0:0:1%1,
>,], sockAddrs=[/,
> /0:0:0:0:0:0:0:1%1:47500, /], discPort=47500, order=271,
> intOrder=145, lastExchangeTime=1476344572001, loc=true,
> ver=1.7.0#20160801-sha1:383273e3, isClient=false]
> WARN : [managers.discovery.GridDiscoveryManager]
> Date=2016/10/13/07/42/52/092|Stopping local node according to configured
> segmentation policy.
> I would like to understand what is default behavior since I didn't
> configure any SegmentationResolvers.
> I can probably set SegmentationPolicy to NOOP to avoid node shutdown, but
> I don't think it's a good idea that node will out of topology for a long
> time.
> Is possible to set time/wait longer until getting SEGMENTED event?
> We are using ignite 1.7.0 and running cluster with 20 nodes.
> Thank you,
> Yitzhak
> --
> Yitzhak Molko

Vladislav Pyatkov

Reply via email to