Ivan Daschinskiy created IGNITE-13491:
-----------------------------------------
Summary: Fix incorrect topology snapshot logger output about
coordinator change.
Key: IGNITE-13491
URL: https://issues.apache.org/jira/browse/IGNITE-13491
Project: Ignite
Issue Type: Bug
Reporter: Ivan Daschinskiy
Currently, logic in
{{org.apache.ignite.internal.managers.discovery.GridDiscoveryManager#topologySnapshotMessage}}
has major drawback, in condition we don't check that failed node with order
less than oldest server node, is actually server node. So we can see invalid
message about coordinator change, event though previous node was a client.
Reproducer:
1. Start server #1
2. Start client
3. Start server #1
4. Stop server #1 and client
We will see in logs of server #2 something like this:
{{[2020-09-29 10:41:25,909][INFO
][disco-event-worker-#150%tcp.TcpDiscoverySpiMBeanTest2%][GridDiscoveryManager]
Coordinator changed [prev=TcpDiscoveryNode
[id=371896fb-f612-4640-bfcd-cef6d2800001,
consistentId=371896fb-f612-4640-bfcd-cef6d2800001, addrs=ArrayList [127.0.0.1],
sockAddrs=HashSet [/127.0.0.1:0], discPort=0, order=2, intOrder=2,
lastExchangeTime=1601365285287, loc=false, ver=2.10.0#20200929-sha1:00000000,
*isClient=true*], cur=TcpDiscoveryNode
[id=9d90f4b0-1374-4147-b7a7-d821f0000002, consistentId=127.0.0.1:47501,
addrs=ArrayList [127.0.0.1], sockAddrs=HashSet [/127.0.0.1:47501],
discPort=47501, order=3, intOrder=3, lastExchangeTime=1601365285900, loc=true,
ver=2.10.0#20200929-sha1:00000000, isClient=false]]}}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
