You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by "Jim Kellerman (JIRA)" <ji...@apache.org> on 2007/12/10 20:19:43 UTC
[jira] Issue Comment Edited: (HADOOP-2392) [hbase]
TestRegionServerExit has new failure mode since HADOOP-2338
[ https://issues.apache.org/jira/browse/HADOOP-2392?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12550144 ]
jimk edited comment on HADOOP-2392 at 12/10/07 11:18 AM:
------------------------------------------------------------------
Failed in nightly build #328
This was a strange failure. The region server serving the meta region aborted as expected, its leas timed out, the master started processing the shutdown by splitting the region server's log and scanning the root region where it discovered that the downed server was serving the meta region, the meta region was reassigned, opened, information was updated in the root region, yet the next root scan got back the old data:
{code}
[junit] 2007-12-10 11:58:52,417 INFO [HMaster] org.apache.hadoop.hbase.HMaster$ProcessRegionOpen.process(HMaster.java:2461): updating row .META.,,1 in table -ROOT-,,0 with startcode 1197287919218 and server 140.211.11.75:39266
[junit] 2007-12-10 11:59:00,221 INFO [HMaster.rootScanner] org.apache.hadoop.hbase.HMaster$BaseScanner.scanRegion(HMaster.java:213): HMaster.rootScanner scanning meta region regionname: -ROOT-,,0, startKey: <>, server: 140.211.11.75:39206}
[junit] 2007-12-10 11:59:00,317 DEBUG [HMaster.rootScanner] org.apache.hadoop.hbase.HMaster$BaseScanner.scanRegion(HMaster.java:249): HMaster.rootScanner regioninfo: {regionname: .META.,,1, startKey: <>, tableDesc: {name: .META., families: {info:={name: info, max versions: 1, compression: NONE, in memory: false, max length: 2147483647, bloom filter: none}}}}, server: 140.211.11.75:39205, startCode: 1197287908915
[junit] 2007-12-10 11:59:00,317 DEBUG [HMaster.rootScanner] org.apache.hadoop.hbase.HMaster$BaseScanner.checkAssigned(HMaster.java:469): Current assignment of .META.,,1 is no good
{code}
As to why the test timed out, one of the region servers failed to exit completely (HRegionServer.run)
was (Author: jimk):
Failed in nightly build #328
> [hbase] TestRegionServerExit has new failure mode since HADOOP-2338
> -------------------------------------------------------------------
>
> Key: HADOOP-2392
> URL: https://issues.apache.org/jira/browse/HADOOP-2392
> Project: Hadoop
> Issue Type: Bug
> Components: contrib/hbase
> Affects Versions: 0.16.0
> Reporter: Jim Kellerman
> Assignee: Jim Kellerman
> Fix For: 0.16.0
>
> Attachments: patch.txt
>
>
> TestRegionServerExit has a new failure mode since HADOOP-2338. It appears that the region server won't exit. Is it possible that a scanner lease is not being cleaned up correctly?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.