Ivan Bessonov created IGNITE-13814:
--------------------------------------
Summary: Long restorePartitionStates triggers FailureHandler on
node startup
Key: IGNITE-13814
URL: https://issues.apache.org/jira/browse/IGNITE-13814
Project: Ignite
Issue Type: Bug
Environment: {noformat}
Thread [name="sys-stripe-4-#5%EPE_CLUSTER_PERF%", id=24, state=WAITING,
blockCnt=4, waitCnt=70836]
at java.base@11.0.8/jdk.internal.misc.Unsafe.park(Native Method)
at
java.base@11.0.8/java.util.concurrent.locks.LockSupport.park(LockSupport.java:323)
at
app//o.a.i.i.util.future.GridFutureAdapter.get0(GridFutureAdapter.java:186)
at
app//o.a.i.i.util.future.GridFutureAdapter.getUninterruptibly(GridFutureAdapter.java:154)
at
app//o.a.i.i.processors.cache.persistence.file.AsyncFileIO.read(AsyncFileIO.java:128)
at
app//o.a.i.i.processors.cache.persistence.file.AbstractFileIO$2.run(AbstractFileIO.java:89)
at
app//o.a.i.i.processors.cache.persistence.file.AbstractFileIO.fully(AbstractFileIO.java:52)
at
app//o.a.i.i.processors.cache.persistence.file.AbstractFileIO.readFully(AbstractFileIO.java:87)
at
app//o.a.i.i.processors.cache.persistence.file.FilePageStore.readWithFailover(FilePageStore.java:794)
at
app//o.a.i.i.processors.cache.persistence.file.FilePageStore.read(FilePageStore.java:418)
at
app//o.a.i.i.processors.cache.persistence.file.FilePageStoreManager.read(FilePageStoreManager.java:519)
at
app//o.a.i.i.processors.cache.persistence.file.FilePageStoreManager.read(FilePageStoreManager.java:503)
at
app//o.a.i.i.processors.cache.persistence.pagemem.PageMemoryImpl.acquirePage(PageMemoryImpl.java:874)
at
app//o.a.i.i.processors.cache.persistence.pagemem.PageMemoryImpl.acquirePage(PageMemoryImpl.java:700)
at
app//o.a.i.i.processors.cache.persistence.pagemem.PageMemoryImpl.acquirePage(PageMemoryImpl.java:689)
at
app//o.a.i.i.processors.cache.persistence.DataStructure.acquirePage(DataStructure.java:157)
at
app//o.a.i.i.processors.cache.persistence.freelist.PagesList.init(PagesList.java:274)
at
app//o.a.i.i.processors.cache.persistence.freelist.AbstractFreeList.<init>(AbstractFreeList.java:390)
at
app//o.a.i.i.processors.cache.persistence.freelist.CacheFreeList.<init>(CacheFreeList.java:57)
at
app//o.a.i.i.processors.cache.persistence.GridCacheOffheapManager$GridCacheDataStore$1.<init>(GridCacheOffheapManager.java:1806)
at
app//o.a.i.i.processors.cache.persistence.GridCacheOffheapManager$GridCacheDataStore.init0(GridCacheOffheapManager.java:1805)
at
app//o.a.i.i.processors.cache.persistence.GridCacheOffheapManager$GridCacheDataStore.init(GridCacheOffheapManager.java:2130)
at
app//o.a.i.i.processors.cache.persistence.GridCacheOffheapManager.restorePartitionStates(GridCacheOffheapManager.java:544)
at
app//o.a.i.i.processors.cache.GridCacheProcessor$CacheRecoveryLifecycle.lambda$restorePartitionStates$0(GridCacheProcessor.java:5253)
at
app//o.a.i.i.processors.cache.GridCacheProcessor$CacheRecoveryLifecycle$$Lambda$633/0x0000000800717040.run(Unknown
Source)
at
app//o.a.i.i.util.StripedExecutor$Stripe.body(StripedExecutor.java:559)
at app//o.a.i.i.util.worker.GridWorker.run(GridWorker.java:119)
at java.base@11.0.8/java.lang.Thread.run(Thread.java:834){noformat}
In this case, warm-up is on, but client also reports this to happen without
warm-up.I don't think that restore partition states should trigger FH. It may
take a lot of time with PDS. Also, why do we run it in striped pool? Let's
imagine two large caches get the same stripe - restore time doubles.
Reporter: Ivan Bessonov
Assignee: Ivan Bessonov
The following would be printed to log:
{noformat}
[2020-10-30T17:32:26,190][WARN ][grid-timeout-worker-#22%EPE_CLUSTER_PERF%][]
Possible failure suppressed accordingly to a configured handler
[hnd=StopNodeOrHaltFailureHandler [tryStop=false, timeout=0,
super=AbstractFailureHandler [ignoredFailureTypes=UnmodifiableSet
[SYSTEM_WORKER_BLOCKED, SYSTEM_CRITICAL_OPERATION_TIMEOUT]]],
failureCtx=FailureContext [type=SYSTEM_WORKER_BLOCKED, err=class
o.a.i.IgniteException: GridWorker [name=sys-stripe-4,
igniteInstanceName=EPE_CLUSTER_PERF, finished=false,
heartbeatTs=1604104192954]]]
org.apache.ignite.IgniteException: GridWorker [name=sys-stripe-4,
igniteInstanceName=EPE_CLUSTER_PERF, finished=false, heartbeatTs=1604104192954]
at
org.apache.ignite.internal.IgnitionEx$IgniteNamedInstance$2.apply(IgnitionEx.java:1859)
[ignite-core-8.7.28.jar:8.7.28]
at
org.apache.ignite.internal.IgnitionEx$IgniteNamedInstance$2.apply(IgnitionEx.java:1854)
[ignite-core-8.7.28.jar:8.7.28]
at
org.apache.ignite.internal.worker.WorkersRegistry.onIdle(WorkersRegistry.java:233)
[ignite-core-8.7.28.jar:8.7.28]
at
org.apache.ignite.internal.util.worker.GridWorker.onIdle(GridWorker.java:296)
[ignite-core-8.7.28.jar:8.7.28]
at
org.apache.ignite.internal.processors.timeout.GridTimeoutProcessor$TimeoutWorker.body(GridTimeoutProcessor.java:265)
[ignite-core-8.7.28.jar:8.7.28]
at
org.apache.ignite.internal.util.worker.GridWorker.run(GridWorker.java:119)
[ignite-core-8.7.28.jar:8.7.28]
at java.lang.Thread.run(Thread.java:834) [?:?]{noformat}
Actual thread dump of affected thread:
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
