[IGNITE-10389] Getting affinity for topology version earlier than affinity is calculated

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 2.6
• Assignee: Unassigned
• Reporter: Arnaud Rivero
• Created: 23/Nov/18 08:21
• Updated: 17/Sep/19 10:35
• Resolved: 17/Sep/19 10:35

Description

When we deploy a new Ignite cluster on YARN and we deploy an Ignite service at the same time (by configuring it in the xml configuration file et putting it in the IGNITE_USERS_LIBS path), we get the following exception (the service name was replaced by XXXXXXXXXX): 

 

 

[14:48:42,004][SEVERE][srvc-deploy-#144][GridServiceProcessor] Error when executing service: XXXXXXXXXX
java.lang.IllegalStateException: Getting affinity for topology version earlier than affinity is calculated [locNode=TcpDiscoveryNode [id=fd770768-189d-4374-82af-2497eff173da, addrs=[<addrs>], sockAddrs=[<sockAddrs>], discPort=47500, order=1, intOrder=1, lastExchangeTime=1542894520229, loc=true, ver=2.6.0#20180710-sha1:669feacc, isClient=false], grp=ignite-sys-cache, topVer=AffinityTopologyVersion [topVer=2, minorTopVer=0], head=AffinityTopologyVersion [topVer=4, minorTopVer=1], history=[AffinityTopologyVersion [topVer=1, minorTopVer=0], AffinityTopologyVersion [topVer=4, minorTopVer=0], AffinityTopologyVersion [topVer=4, minorTopVer=1]]]
at org.apache.ignite.internal.processors.affinity.GridAffinityAssignmentCache.cachedAffinity(GridAffinityAssignmentCache.java:632)
at org.apache.ignite.internal.processors.affinity.GridAffinityAssignmentCache.nodes(GridAffinityAssignmentCache.java:532)
at org.apache.ignite.internal.processors.cache.GridCacheAffinityManager.nodesByPartition(GridCacheAffinityManager.java:225)
at org.apache.ignite.internal.processors.cache.GridCacheAffinityManager.primaryByPartition(GridCacheAffinityManager.java:261)
at org.apache.ignite.internal.processors.cache.GridCacheAffinityManager.primaryByKey(GridCacheAffinityManager.java:252)
at org.apache.ignite.internal.processors.cache.GridCacheAffinityManager.primaryByKey(GridCacheAffinityManager.java:276)
at org.apache.ignite.internal.processors.service.GridServiceProcessor$TopologyListener$1.run0(GridServiceProcessor.java:1828)
at org.apache.ignite.internal.processors.service.GridServiceProcessor$DepRunnable.run(GridServiceProcessor.java:2015)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:745)

 

We found this previous issue and we wondered if we didn't meet an edge case of it : 

IGNITE-1171

 

Issue Links

• duplicates: IGNITE-11465 Multiple client leave/join events may wipe affinity assignment history and cause transactions fail
• is related to: IGNITE-12014 Getting affinity for topology version earlier than affinity is calculated for system cache