hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jim Kellerman (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-2324) [hbase] Fix assertion failures in TestTableMapReduce
Date Mon, 10 Dec 2007 18:45:43 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-2324?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12550148
] 

Jim Kellerman commented on HADOOP-2324:
---------------------------------------

Failed in nightly build #328

The weird thing about this failure is that the master updates the meta to indicate that the
child split regions are being served, the client can find the children in the meta table,
but the meta table scanner doesn't and proceeds to assign the child regions again:

{code}
    [junit] 2007-12-10 12:22:53,353 INFO  [HMaster] org.apache.hadoop.hbase.HMaster$ProcessRegionOpen.process(HMaster.java:2461):
updating row mrtest,kps~`@#$%^&*()-_+=:;',.<>/?[]{}|,1197289367489 in table .META.,,1
with startcode 1197289296412 and server 140.211.11.75:35921

    [junit] 2007-12-10 12:22:53,355 INFO  [HMaster] org.apache.hadoop.hbase.HMaster$ProcessRegionOpen.process(HMaster.java:2461):
updating row mrtest,eya~`@#$%^&*()-_+=:;',.<>/?[]{}|,1197289367489 in table .META.,,1
with startcode 1197289296412 and server 140.211.11.75:35921

    [junit] 2007-12-10 12:23:02,726 INFO  [HMaster.metaScanner] org.apache.hadoop.hbase.HMaster$BaseScanner.scanRegion(HMaster.java:213):
HMaster.metaScanner scanning meta region regionname: .META.,,1, startKey: <>, server:
140.211.11.75:35921}

    [junit] 2007-12-10 12:23:02,738 DEBUG [HMaster.metaScanner] org.apache.hadoop.hbase.HMaster$BaseScanner.scanRegion(HMaster.java:249):
HMaster.metaScanner regioninfo: {regionname: mrtest,eya~`@#$%^&*()-_+=:;',.<>/?[]{}|,1197289367489,
startKey: <eya~`@#$%^&*()-_+=:;',.<>/?[]{}|>, tableDesc: {name: mrtest, families:
{contents:={name: contents, max versions: 3, compression: NONE, in memory: false, max length:
2147483647, bloom filter: none}, text:={name: text, max versions: 3, compression: NONE, in
memory: false, max length: 2147483647, bloom filter: none}}}}, server: , startCode: -1
    [junit] 2007-12-10 12:23:02,738 DEBUG [HMaster.metaScanner] org.apache.hadoop.hbase.HMaster$BaseScanner.checkAssigned(HMaster.java:469):
Current assignment of mrtest,eya~`@#$%^&*()-_+=:;',.<>/?[]{}|,1197289367489 is no
good
    [junit] 2007-12-10 12:23:02,739 DEBUG [HMaster.metaScanner] org.apache.hadoop.hbase.HMaster$BaseScanner.scanRegion(HMaster.java:249):
HMaster.metaScanner regioninfo: {regionname: mrtest,kps~`@#$%^&*()-_+=:;',.<>/?[]{}|,1197289367489,
startKey: <kps~`@#$%^&*()-_+=:;',.<>/?[]{}|>, tableDesc: {name: mrtest, families:
{contents:={name: contents, max versions: 3, compression: NONE, in memory: false, max length:
2147483647, bloom filter: none}, text:={name: text, max versions: 3, compression: NONE, in
memory: false, max length: 2147483647, bloom filter: none}}}}, server: , startCode: -1
    [junit] 2007-12-10 12:23:02,739 DEBUG [HMaster.metaScanner] org.apache.hadoop.hbase.HMaster$BaseScanner.checkAssigned(HMaster.java:469):
Current assignment of mrtest,kps~`@#$%^&*()-_+=:;',.<>/?[]{}|,1197289367489 is no
good
{code}

> [hbase] Fix assertion failures in TestTableMapReduce
> ----------------------------------------------------
>
>                 Key: HADOOP-2324
>                 URL: https://issues.apache.org/jira/browse/HADOOP-2324
>             Project: Hadoop
>          Issue Type: Bug
>          Components: contrib/hbase
>    Affects Versions: 0.16.0
>            Reporter: Jim Kellerman
>            Assignee: Jim Kellerman
>            Priority: Minor
>             Fix For: 0.16.0
>
>
> TestTableMapReduce consists of two tests: a very simple one that verifies that it is
working, and a larger test that verifies that it works when the table consists of multiple
regions.
> There are two failure modes:
> 1) an assertion failure in the verify method
> 2) the test times out because something does not shut down cleanly.
> to figure out what is causing these failures, the first step will be to do copious logging
in the test. Once the root cause is determined, we can address it.
> Unfortunately, some of the testing will have to be done via Hudson, because that is the
only environment in which it fails. We routinely test under Windows, Fedora Core 6, MacOsX,
and Ubuntu. Windows and Ubuntu have both single and dual processor configurations. FC6 and
MacOS run on dual core machines.
> Only Hudson seems crippled enough to demonstrate these problems. (A good thing since
they are real problems)

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message