[
https://issues.apache.org/jira/browse/DERBY-5185?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dag H. Wanvik updated DERBY-5185:
---------------------------------
Attachment: derby-5185-2a.stat
derby-5185-2a.diff
Scenario: A thread has seen an interrupt and is getting ready to resurrect the
channel in "recoverContainerAfterInterrupt". Before it starts doing this,
however, it waits for any other threads currently doing IO on this container to
"hit the wall" and start waiting for the current thread to do the recovery. The
current thread knows its free to proceed when the counter "threadsInPageIO"
reaches 0. In this case, it has given up waiting for the counter the reach 0
and throws FILE_IO_INTERRUPTED. (I believe "threadsInPageIO" should have been 0
here but is not for some reason, see below). In throwing, it neglects to reset
state variable "restoreChannelInProgress" which makes the (next) thread coming
along, which we see hanging in this issue get stuck when trying to enter the IO
code in the "gain entry" section.
Attaching a patch derby-5185-2a, which fixes state invariant maintenance when
throwing FILE_IO_INTERRUPTED. It also adds a maximum number of retries for the
readPage code abd fixes some cases whereby the state variable "threadsInPageIO"
could risk not being properly update when exceptions would get thrown. This may
be the underlying reason for what we see here.
Ran regressions OK on Solaris/JDK6 and Debian/JDK7.
> store/rollForwardRecovery.sql stuck in
> RAFContainer4.recoverContainerAfterInterrupt() during shutdown
> -----------------------------------------------------------------------------------------------------
>
> Key: DERBY-5185
> URL: https://issues.apache.org/jira/browse/DERBY-5185
> Project: Derby
> Issue Type: Bug
> Components: Store
> Affects Versions: 10.9.0.0
> Environment: Derby 10.9.0.0 alpha - (1090406)
> Oracle Enterprise Linux 6.0
> Linux 2.6.32-100.28.9.el6.x86_64 #1 SMP Wed Mar 16 19:24:16 EDT 2011 x86_64
> x86_64 x86_64 GNU/Linux
> java version "1.7.0-ea"
> Java(TM) SE Runtime Environment (build 1.7.0-ea-b135)
> Java HotSpot(TM) Client VM (build 21.0-b05, mixed mode, sharing)
> Reporter: Knut Anders Hatlen
> Attachments: derby-5185-1a.diff, derby-5185-2a.diff,
> derby-5185-2a.stat, stack.txt
>
>
> I have a derbyall that has been running for more than two days now. It seems
> to be stuck in the store/rollForwardRecovery.sql test while the engine is
> shutting down.
> Here's the stack trace for the daemon thread that's stuck:
> "derby.rawStoreDaemon" daemon prio=10 tid=0xf3e7dc00 nid=0x3505 waiting on
> condition [0xf4066000]
> java.lang.Thread.State: TIMED_WAITING (sleeping)
> at java.lang.Thread.sleep(Native Method)
> at
> org.apache.derby.impl.store.raw.data.RAFContainer4.recoverContainerAfterInterrupt(Unknown
> Source)
> at
> org.apache.derby.impl.store.raw.data.RAFContainer4.readPage(Unknown Source)
> at
> org.apache.derby.impl.store.raw.data.RAFContainer4.readPage(Unknown Source)
> at org.apache.derby.impl.store.raw.data.CachedPage.readPage(Unknown
> Source)
> at
> org.apache.derby.impl.store.raw.data.CachedPage.setIdentity(Unknown Source)
> at org.apache.derby.impl.services.cache.ConcurrentCache.find(Unknown
> Source)
> at
> org.apache.derby.impl.store.raw.data.FileContainer.getAllocPage(Unknown
> Source)
> at
> org.apache.derby.impl.store.raw.data.BaseContainer.getAllocPage(Unknown
> Source)
> at
> org.apache.derby.impl.store.raw.data.BaseContainerHandle.getAllocPage(Unknown
> Source)
> at
> org.apache.derby.impl.store.raw.data.FileContainer.deallocatePagenum(Unknown
> Source)
> - locked <0xc5adbce8> (a
> org.apache.derby.impl.store.raw.data.AllocationCache)
> at
> org.apache.derby.impl.store.raw.data.FileContainer.deallocatePage(Unknown
> Source)
> at
> org.apache.derby.impl.store.raw.data.BaseContainer.removePage(Unknown Source)
> at
> org.apache.derby.impl.store.raw.data.BaseContainerHandle.removePage(Unknown
> Source)
> at
> org.apache.derby.impl.store.access.heap.HeapController.removePage(Unknown
> Source)
> at
> org.apache.derby.impl.store.access.heap.HeapPostCommit.purgeCommittedDeletes(Unknown
> Source)
> at
> org.apache.derby.impl.store.access.heap.HeapPostCommit.performWork(Unknown
> Source)
> at
> org.apache.derby.impl.services.daemon.BasicDaemon.serviceClient(Unknown
> Source)
> at org.apache.derby.impl.services.daemon.BasicDaemon.work(Unknown
> Source)
> at org.apache.derby.impl.services.daemon.BasicDaemon.run(Unknown
> Source)
> at java.lang.Thread.run(Thread.java:722)
> And here's the stack trace for the main thread, which is waiting for the
> daemon thread to stop:
> "main" prio=10 tid=0xf6c05c00 nid=0x34e5 in Object.wait() [0xf6dbe000]
> java.lang.Thread.State: WAITING (on object monitor)
> at java.lang.Object.wait(Native Method)
> - waiting on <0xc5ac5760> (a
> org.apache.derby.impl.services.daemon.BasicDaemon)
> at java.lang.Object.wait(Object.java:504)
> at org.apache.derby.impl.services.daemon.BasicDaemon.pause(Unknown
> Source)
> - locked <0xc5ac5760> (a
> org.apache.derby.impl.services.daemon.BasicDaemon)
> at org.apache.derby.impl.services.daemon.BasicDaemon.stop(Unknown
> Source)
> at org.apache.derby.impl.store.raw.RawStore.stop(Unknown Source)
> at org.apache.derby.impl.services.monitor.TopService.stop(Unknown
> Source)
> at org.apache.derby.impl.services.monitor.TopService.shutdown(Unknown
> Source)
> at
> org.apache.derby.impl.services.monitor.BaseMonitor.shutdown(Unknown Source)
> at
> org.apache.derby.impl.services.monitor.BaseMonitor.shutdown(Unknown Source)
> at org.apache.derby.jdbc.InternalDriver.connect(Unknown Source)
> at org.apache.derby.jdbc.AutoloadedDriver.connect(Unknown Source)
> at java.sql.DriverManager.getConnection(DriverManager.java:620)
> at java.sql.DriverManager.getConnection(DriverManager.java:222)
> at org.apache.derby.impl.tools.ij.utilMain.cleanupGo(Unknown Source)
> at org.apache.derby.impl.tools.ij.utilMain.go(Unknown Source)
> at org.apache.derby.impl.tools.ij.Main.go(Unknown Source)
> at org.apache.derby.impl.tools.ij.Main.mainCore(Unknown Source)
> at org.apache.derby.impl.tools.ij.Main.main(Unknown Source)
> at org.apache.derby.tools.ij.main(Unknown Source)
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
