FAQ
Disabling a table can return too soon if one region is pending open
-------------------------------------------------------------------

Key: HBASE-2067
URL: https://issues.apache.org/jira/browse/HBASE-2067
Project: Hadoop HBase
Issue Type: Sub-task
Affects Versions: 0.20.2
Reporter: Jean-Daniel Cryans
Assignee: Jean-Daniel Cryans
Fix For: 0.20.3, 0.21.0


Still using my test I was grepping for "Duplicate assignment" and found that:

{code}
2009-12-22 01:43:31,450 DEBUG [IPC Server handler 3 on 60000] master.ChangeTableState(70): Processing unserved regions
2009-12-22 01:43:31,450 INFO [RegionServer:2.worker] regionserver.HRegion(343): region table8,,1261474997182/1735800584 available; sequence id is 0
2009-12-22 01:43:31,451 DEBUG [IPC Server handler 3 on 60000] master.ChangeTableState(102): Removed server and startcode from row and set online=false: table8,,1261474997182
2009-12-22 01:43:31,451 DEBUG [IPC Server handler 3 on 60000] master.ChangeTableState(120): Processing regions currently being served2009-12-22 01:43:31,451 DEBUG [main] client.HConnectionManager$TableServers(776): Cache hit for row <> in tableName .META.: location server 192.168.1.103:54769, location region
name .META.,,1
2009-12-22 01:43:31,456 DEBUG [main] client.HConnectionManager$TableServers(554): Rowscanned=1, rowsOffline=1
2009-12-22 01:43:31,456 INFO [main] client.HBaseAdmin(414): Disabled table8
2009-12-22 01:43:31,459 DEBUG [IPC Server handler 1 on 60000] regionserver.HRegion(2003): DELETING region hdfs://localhost:54729/user/jdcryans/table8/1735800584
2009-12-22 01:43:31,461 INFO [IPC Server handler 1 on 60000] master.HMaster(807): deleted table: table8
2009-12-22 01:43:31,464 INFO [main] client.HBaseAdmin(302): Deleted table8
...
2009-12-22 01:43:31,541 INFO [IPC Server handler 2 on 60000] master.ServerManager(421): Processing MSG_REPORT_PROCESS_OPEN: table8,,1261474997182 from 192.168.1.103,54771,1261
474989274; 1 of 2
2009-12-22 01:43:31,541 INFO [IPC Server handler 2 on 60000] master.ServerManager(421): Processing MSG_REPORT_OPEN: table8,,1261474997182 from 192.168.1.103,54771,1261474989274; 2 of 2
2009-12-22 01:43:31,541 DEBUG [IPC Server handler 2 on 60000] master.ServerManager(567): region server 192.168.1.103:54771 should not have opened region table8,,1261474997182
2009-12-22 01:43:31,542 INFO [RegionServer:2] regionserver.HRegionServer(475): MSG_REGION_CLOSE_WITHOUT_REPORT: table8,,1261474997182: Duplicate assignment
2009-12-22 01:43:31,542 INFO [RegionServer:2.worker] regionserver.HRegionServer$Worker(1332): Worker: MSG_REGION_CLOSE_WITHOUT_REPORT: table8,,1261474997182: Duplicate assignm
ent
{code}

So the region is marked as offline while it's pending open. The thread in HCM.TableServer sees that as if the region was closed but it's not and it's reporting just a little bit after.

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

Search Discussions

  • Jean-Daniel Cryans (JIRA) at Dec 28, 2009 at 6:19 pm
    [ https://issues.apache.org/jira/browse/HBASE-2067?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

    Jean-Daniel Cryans resolved HBASE-2067.
    ---------------------------------------

    Resolution: Duplicate

    Fixed by HBASE-2065
    Disabling a table can return too soon if one region is pending open
    -------------------------------------------------------------------

    Key: HBASE-2067
    URL: https://issues.apache.org/jira/browse/HBASE-2067
    Project: Hadoop HBase
    Issue Type: Sub-task
    Affects Versions: 0.20.2
    Reporter: Jean-Daniel Cryans
    Assignee: Jean-Daniel Cryans
    Fix For: 0.20.3, 0.21.0


    Still using my test I was grepping for "Duplicate assignment" and found that:
    {code}
    2009-12-22 01:43:31,450 DEBUG [IPC Server handler 3 on 60000] master.ChangeTableState(70): Processing unserved regions
    2009-12-22 01:43:31,450 INFO [RegionServer:2.worker] regionserver.HRegion(343): region table8,,1261474997182/1735800584 available; sequence id is 0
    2009-12-22 01:43:31,451 DEBUG [IPC Server handler 3 on 60000] master.ChangeTableState(102): Removed server and startcode from row and set online=false: table8,,1261474997182
    2009-12-22 01:43:31,451 DEBUG [IPC Server handler 3 on 60000] master.ChangeTableState(120): Processing regions currently being served2009-12-22 01:43:31,451 DEBUG [main] client.HConnectionManager$TableServers(776): Cache hit for row <> in tableName .META.: location server 192.168.1.103:54769, location region
    name .META.,,1
    2009-12-22 01:43:31,456 DEBUG [main] client.HConnectionManager$TableServers(554): Rowscanned=1, rowsOffline=1
    2009-12-22 01:43:31,456 INFO [main] client.HBaseAdmin(414): Disabled table8
    2009-12-22 01:43:31,459 DEBUG [IPC Server handler 1 on 60000] regionserver.HRegion(2003): DELETING region hdfs://localhost:54729/user/jdcryans/table8/1735800584
    2009-12-22 01:43:31,461 INFO [IPC Server handler 1 on 60000] master.HMaster(807): deleted table: table8
    2009-12-22 01:43:31,464 INFO [main] client.HBaseAdmin(302): Deleted table8
    ...
    2009-12-22 01:43:31,541 INFO [IPC Server handler 2 on 60000] master.ServerManager(421): Processing MSG_REPORT_PROCESS_OPEN: table8,,1261474997182 from 192.168.1.103,54771,1261
    474989274; 1 of 2
    2009-12-22 01:43:31,541 INFO [IPC Server handler 2 on 60000] master.ServerManager(421): Processing MSG_REPORT_OPEN: table8,,1261474997182 from 192.168.1.103,54771,1261474989274; 2 of 2
    2009-12-22 01:43:31,541 DEBUG [IPC Server handler 2 on 60000] master.ServerManager(567): region server 192.168.1.103:54771 should not have opened region table8,,1261474997182
    2009-12-22 01:43:31,542 INFO [RegionServer:2] regionserver.HRegionServer(475): MSG_REGION_CLOSE_WITHOUT_REPORT: table8,,1261474997182: Duplicate assignment
    2009-12-22 01:43:31,542 INFO [RegionServer:2.worker] regionserver.HRegionServer$Worker(1332): Worker: MSG_REGION_CLOSE_WITHOUT_REPORT: table8,,1261474997182: Duplicate assignm
    ent
    {code}
    So the region is marked as offline while it's pending open. The thread in HCM.TableServer sees that as if the region was closed but it's not and it's reporting just a little bit after.
    --
    This message is automatically generated by JIRA.
    -
    You can reply to this email to add a comment to the issue online.

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupdev @
categorieshbase, hadoop
postedDec 22, '09 at 10:08a
activeDec 28, '09 at 6:19p
posts2
users1
websitehbase.apache.org

1 user in discussion

Jean-Daniel Cryans (JIRA): 2 posts

People

Translate

site design / logo © 2022 Grokbase