[ https://issues.apache.org/jira/browse/IGNITE-21010?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17800005#comment-17800005 ]
Roman Puchkovskiy commented on IGNITE-21010: -------------------------------------------- A bunch of issues were created after investigation: IGNITE-20978, IGNITE-21079, IGNITE-21068 > Cluster loses connectivity after some time > ------------------------------------------ > > Key: IGNITE-21010 > URL: https://issues.apache.org/jira/browse/IGNITE-21010 > Project: Ignite > Issue Type: Bug > Reporter: Roman Puchkovskiy > Assignee: Roman Puchkovskiy > Priority: Major > Labels: ignite-3 > > When running ItSqlLogicTest (where the cluster consists of 2 nodes), > sometimes nodes lose each other (i.e. both think that anoher node has left > the physical topology). > This is because ScaleCube cannot perform ping roundtrip (send ping request + > receive ping response) in 200ms, and this state holds for 10 seconds. > Either this is because the timeout of 200ms is too low, or there is something > summing up (like a leak, or a message that blocks the processing for too > long). This has to be investigated. -- This message was sent by Atlassian Jira (v8.20.10#820010)