[
https://issues.apache.org/jira/browse/HBASE-20081?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16376401#comment-16376401
]
stack commented on HBASE-20081:
-------------------------------
The link doesn't work.
How often does the test fail?
Why would a region (not meta) not being online hold up shutdown?
bq. Since DNS misconfiguration was very unlikely on Apache Jenkins nodes, the
above should not have been logged.
?
bq. java.io.IOException: connection is closed....The above was possibly related
to the lost region server.
?
The server connection is closed on shutdown. If a catalog janitor running, it
will get connection closed.
What does any of the above have to do w/ a hung shutdown?
Did the test timeout? Isn't there a thread dump?
> TestDisableTableProcedure sometimes hung in MiniHBaseCluster#waitUntilShutDown
> ------------------------------------------------------------------------------
>
> Key: HBASE-20081
> URL: https://issues.apache.org/jira/browse/HBASE-20081
> Project: HBase
> Issue Type: Test
> Reporter: Ted Yu
> Priority: Major
>
> https://builds.apache.org/job/HBase-2.0-hadoop3-tests/lastCompletedBuild/org.apache.hbase$hbase-server/testReport/org.apache.hadoop.hbase.master.procedure/TestDisableTableProcedure/org_apache_hadoop_hbase_master_procedure_TestDisableTableProcedure/
> was one recent occurrence.
> I noticed two things in test output:
> {code}
> 2018-02-25 18:12:45,053 WARN [Time-limited test-EventThread]
> master.RegionServerTracker(136): asf912.gq1.ygridcore.net,45649,1519582305777
> is not online or isn't known to the master.The latter could be caused by a
> DNS misconfiguration.
> {code}
> Since DNS misconfiguration was very unlikely on Apache Jenkins nodes, the
> above should not have been logged.
> {code}
> 2018-02-25 18:16:51,531 WARN [master/asf912:0.Chore.1]
> master.CatalogJanitor(127): Failed scan of catalog table
> java.io.IOException: connection is closed
> at
> org.apache.hadoop.hbase.MetaTableAccessor.getMetaHTable(MetaTableAccessor.java:263)
> at
> org.apache.hadoop.hbase.MetaTableAccessor.scanMeta(MetaTableAccessor.java:761)
> at
> org.apache.hadoop.hbase.MetaTableAccessor.scanMeta(MetaTableAccessor.java:680)
> at
> org.apache.hadoop.hbase.MetaTableAccessor.scanMetaForTableRegions(MetaTableAccessor.java:675)
> at
> org.apache.hadoop.hbase.master.CatalogJanitor.getMergedRegionsAndSplitParents(CatalogJanitor.java:188)
> at
> org.apache.hadoop.hbase.master.CatalogJanitor.getMergedRegionsAndSplitParents(CatalogJanitor.java:140)
> at
> org.apache.hadoop.hbase.master.CatalogJanitor.scan(CatalogJanitor.java:246)
> at
> org.apache.hadoop.hbase.master.CatalogJanitor.chore(CatalogJanitor.java:119)
> at org.apache.hadoop.hbase.ScheduledChore.run(ScheduledChore.java:186)
> {code}
> The above was possibly related to the lost region server.
> I searched test output of successful run where none of the above two can be
> seen.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
