You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Alexander Belyak (JIRA)" <ji...@apache.org> on 2017/06/08 08:21:18 UTC
[jira] [Commented] (IGNITE-5445) ServerImpl can't process
NodeFailedMessage about itself
[ https://issues.apache.org/jira/browse/IGNITE-5445?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16042398#comment-16042398 ]
Alexander Belyak commented on IGNITE-5445:
------------------------------------------
In log's I see:
TcpDiscoverSpi's message worker thread failed abnormally. Stopping the node in order to prevent
cluster wide instability.
java.lang.AssertionError: null
at org.apache.ignite.spi.discovery.tcp.internal.TcpDiscoveryNodesRing.removeNode(TcpDiscoveryNodesRing.java:353)
> ServerImpl can't process NodeFailedMessage about itself
> -------------------------------------------------------
>
> Key: IGNITE-5445
> URL: https://issues.apache.org/jira/browse/IGNITE-5445
> Project: Ignite
> Issue Type: Bug
> Components: general
> Affects Versions: 1.9
> Reporter: Alexander Belyak
> Priority: Minor
>
> If for some reason (GC pause or heavy load) node get NodeLeft(FAILED) message about itself - it can't correctly handle it, because it call TcpDiscoveryNodesRing.removeNode with local node id and get assertion error.
> I think - node should correctly determine such event and throw something like "segmented" event and so on.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)