[IGNITE-14138] Historical rebalance kills cluster

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.10
• Release Note: Fixed stopping node in some cases when historical rebalance could not find reserved WAL segments. Now the node refuses from supplying partitions historically.
• Ignite Flags: Docs Required, Release Notes Required
• Assignee: Vladislav Pyatkov
• Reporter: Vladislav Pyatkov
• Reviewer: Slava Koptilin
• Created: 08/Feb/21 11:35
• Updated: 10/Mar/21 14:38
• Resolved: 12/Feb/21 10:14

Description

[2021-01-12T05:11:02,142][ERROR][rebalance-#508%---%][] Critical system error detected. Will be handled accordingly to configured handler [hnd=StopNodeOrHaltFailureHandler [tryStop=false, timeout=0, super=AbstractFailureHandler [ignoredFailureTypes=UnmodifiableSet [SYSTEM_WORKER_BLOCKED, SYSTEM_CRITICAL_OPERATION_TIMEOUT]]], failureCtx=FailureContext [type=CRITICAL_ERROR, err=class o.a.i.IgniteCheckedException: Failed to continue supplying [grp=SQL_USAGES_EPE, demander=48254935-7aa9-4ab5-b398-fdaec334fab7, topVer=AffinityTopologyVersion [topVer=3, minorTopVer=1]]]]
org.apache.ignite.IgniteCheckedException: Failed to continue supplying [grp=SQL_1, demander=48254935-7aa9-4ab5-b398-fdaec334fab7, topVer=AffinityTopologyVersion [topVer=3, minorTopVer=1]]
	at org.apache.ignite.internal.processors.cache.distributed.dht.preloader.GridDhtPartitionSupplier.handleDemandMessage(GridDhtPartitionSupplier.java:571) [ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.distributed.dht.preloader.GridDhtPreloader.handleDemandMessage(GridDhtPreloader.java:398) [ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.GridCachePartitionExchangeManager$5.apply(GridCachePartitionExchangeManager.java:489) [ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.GridCachePartitionExchangeManager$5.apply(GridCachePartitionExchangeManager.java:474) [ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.GridCacheIoManager.processMessage(GridCacheIoManager.java:1142) [ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.GridCacheIoManager.onMessage0(GridCacheIoManager.java:591) [ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.GridCacheIoManager.access$800(GridCacheIoManager.java:109) [ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.GridCacheIoManager$OrderedMessageListener.onMessage(GridCacheIoManager.java:1707) [ignite-core.jar]
	at org.apache.ignite.internal.managers.communication.GridIoManager.invokeListener(GridIoManager.java:1721) [ignite-core.jar]
	at org.apache.ignite.internal.managers.communication.GridIoManager.access$4300(GridIoManager.java:157) [ignite-core.jar]
	at org.apache.ignite.internal.managers.communication.GridIoManager$GridCommunicationMessageSet.unwind(GridIoManager.java:3011) [ignite-core.jar]
	at org.apache.ignite.internal.managers.communication.GridIoManager.unwindMessageSet(GridIoManager.java:1662) [ignite-core.jar]
	at org.apache.ignite.internal.managers.communication.GridIoManager.access$4900(GridIoManager.java:157) [ignite-core.jar]
	at org.apache.ignite.internal.managers.communication.GridIoManager$9.run(GridIoManager.java:1629) [ignite-core.jar]
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128) [?:?]
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628) [?:?]
	at java.lang.Thread.run(Thread.java:834) [?:?]
Caused by: org.apache.ignite.IgniteCheckedException: Could not find start pointer for partition [part=4, partCntrSince=1115]
	at org.apache.ignite.internal.processors.cache.persistence.checkpoint.CheckpointHistory.searchEarliestWalPointer(CheckpointHistory.java:557) ~[ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.persistence.GridCacheOffheapManager.historicalIterator(GridCacheOffheapManager.java:1121) ~[ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.IgniteCacheOffheapManagerImpl.rebalanceIterator(IgniteCacheOffheapManagerImpl.java:1195) ~[ignite-core.jar]
	at org.apache.ignite.internal.processors.cache.distributed.dht.preloader.GridDhtPartitionSupplier.handleDemandMessage(GridDhtPartitionSupplier.java:322) ~[ignite-core.jar]
	... 16 more

I believe that it should throw IgniteHistoricalIteratorException instead of IgniteCheckedException, so it can be properly handled and rebalance can move to the full rebalance instead of killing nodes

Issue Links

• links to: GitHub Pull Request #8769