[
https://issues.apache.org/jira/browse/SOLR-8335?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15027024#comment-15027024
]
Uwe Schindler commented on SOLR-8335:
-------------------------------------
This was just broken behaviour in 4.x. The lock file should not have a prefix,
because this would have no effect at all. Every Solr node would produce a new
lock file name so locking would have no effect :-)
I was expecting that you had the 4.x deprecated setting to forcefully remove
the lock file on startup, which is no longer supported in 5.x. You have to
delete the file on your own.
> HdfsLockFactory does not allow core to come up after a node was killed
> ----------------------------------------------------------------------
>
> Key: SOLR-8335
> URL: https://issues.apache.org/jira/browse/SOLR-8335
> Project: Solr
> Issue Type: Bug
> Affects Versions: 5.0, 5.1, 5.2, 5.2.1, 5.3, 5.3.1
> Reporter: Varun Thacker
>
> When using HdfsLockFactory if a node gets killed instead of a graceful
> shutdown the write.lock file remains in HDFS . The next time you start the
> node the core doesn't load up because of LockObtainFailedException .
> I was able to reproduce this in all 5.x versions of Solr . The problem wasn't
> there when I tested it in 4.10.4
> Steps to reproduce this on 5.x
> 1. Create directory in HDFS : {{bin/hdfs dfs -mkdir /solr}}
> 2. Start Solr: {{bin/solr start -Dsolr.directoryFactory=HdfsDirectoryFactory
> -Dsolr.lock.type=hdfs -Dsolr.data.dir=hdfs://localhost:9000/solr
> -Dsolr.updatelog=hdfs://localhost:9000/solr}}
> 3. Create core: {{./bin/solr create -c test -n data_driven}}
> 4. Kill solr
> 5. The lock file is there in HDFS and is called {{write.lock}}
> 6. Start Solr again and you get a stack trace like this:
> {code}
> 2015-11-23 13:28:04.287 ERROR (coreLoadExecutor-6-thread-1) [ x:test]
> o.a.s.c.CoreContainer Error creating core [test]: Index locked for write for
> core 'test'. Solr now longer supports forceful unlocking via
> 'unlockOnStartup'. Please verify locks manually!
> org.apache.solr.common.SolrException: Index locked for write for core 'test'.
> Solr now longer supports forceful unlocking via 'unlockOnStartup'. Please
> verify locks manually!
> at org.apache.solr.core.SolrCore.<init>(SolrCore.java:820)
> at org.apache.solr.core.SolrCore.<init>(SolrCore.java:659)
> at org.apache.solr.core.CoreContainer.create(CoreContainer.java:723)
> at org.apache.solr.core.CoreContainer$1.call(CoreContainer.java:443)
> at org.apache.solr.core.CoreContainer$1.call(CoreContainer.java:434)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> org.apache.solr.common.util.ExecutorUtil$MDCAwareThreadPoolExecutor$1.run(ExecutorUtil.java:210)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> Caused by: org.apache.lucene.store.LockObtainFailedException: Index locked
> for write for core 'test'. Solr now longer supports forceful unlocking via
> 'unlockOnStartup'. Please verify locks manually!
> at org.apache.solr.core.SolrCore.initIndex(SolrCore.java:528)
> at org.apache.solr.core.SolrCore.<init>(SolrCore.java:761)
> ... 9 more
> 2015-11-23 13:28:04.289 ERROR (coreContainerWorkExecutor-2-thread-1) [ ]
> o.a.s.c.CoreContainer Error waiting for SolrCore to be created
> java.util.concurrent.ExecutionException:
> org.apache.solr.common.SolrException: Unable to create core [test]
> at java.util.concurrent.FutureTask.report(FutureTask.java:122)
> at java.util.concurrent.FutureTask.get(FutureTask.java:192)
> at org.apache.solr.core.CoreContainer$2.run(CoreContainer.java:472)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> org.apache.solr.common.util.ExecutorUtil$MDCAwareThreadPoolExecutor$1.run(ExecutorUtil.java:210)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> Caused by: org.apache.solr.common.SolrException: Unable to create core [test]
> at org.apache.solr.core.CoreContainer.create(CoreContainer.java:737)
> at org.apache.solr.core.CoreContainer$1.call(CoreContainer.java:443)
> at org.apache.solr.core.CoreContainer$1.call(CoreContainer.java:434)
> ... 5 more
> Caused by: org.apache.solr.common.SolrException: Index locked for write for
> core 'test'. Solr now longer supports forceful unlocking via
> 'unlockOnStartup'. Please verify locks manually!
> at org.apache.solr.core.SolrCore.<init>(SolrCore.java:820)
> at org.apache.solr.core.SolrCore.<init>(SolrCore.java:659)
> at org.apache.solr.core.CoreContainer.create(CoreContainer.java:723)
> ... 7 more
> Caused by: org.apache.lucene.store.LockObtainFailedException: Index locked
> for write for core 'test'. Solr now longer supports forceful unlocking via
> 'unlockOnStartup'. Please verify locks manually!
> at org.apache.solr.core.SolrCore.initIndex(SolrCore.java:528)
> at org.apache.solr.core.SolrCore.<init>(SolrCore.java:761)
> ... 9 more
> {code}
> In 4.10.4 I saw these two differences
> 1. The lock file name was different . It's something like :
> {{/solr/index/HdfsDirectory@46ad6bd3
> lockFactory=org.apache.solr.store.hdfs.hdfslockfact...@4b44b5f6-write.lock}}
> 2. When the node is started again after it was killed , it loaded up the core
> just fine but there were two lock files in hdfs now . 4b44b5f6-write.lock is
> the latest one
> {code}
> /solr/index/HdfsDirectory@46ad6bd3
> lockFactory=org.apache.solr.store.hdfs.hdfslockfact...@4b44b5f6-write.lock
> /solr/index/HdfsDirectory@52959724
> lockFactory=org.apache.solr.store.hdfs.hdfslockfact...@9d59d3f-write.lock
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
