You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "ryan rawson (JIRA)" <ji...@apache.org> on 2009/08/26 03:03:59 UTC

[jira] Created: (HBASE-1795) log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'

log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'
------------------------------------------------------------------------------------------

                 Key: HBASE-1795
                 URL: https://issues.apache.org/jira/browse/HBASE-1795
             Project: Hadoop HBase
          Issue Type: Bug
    Affects Versions: 0.20.0, 0.21.0
            Reporter: ryan rawson
             Fix For: 0.20.1, 0.21.0


during log recovery, we dont reset Store.maxSeqId, thus new log entries are stamped starting off from the old files.  This can cause a problem if we fail and recover again, since the new mutations are deemed "old" and shouldnt be applied in a subsequent recovery scenario.

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

[jira] Assigned: (HBASE-1795) log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'

Posted by "stack (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-1795?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

stack reassigned HBASE-1795:
----------------------------

    Assignee: ryan rawson

> log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'
> ------------------------------------------------------------------------------------------
>
>                 Key: HBASE-1795
>                 URL: https://issues.apache.org/jira/browse/HBASE-1795
>             Project: Hadoop HBase
>          Issue Type: Bug
>    Affects Versions: 0.20.0, 0.21.0
>            Reporter: ryan rawson
>            Assignee: ryan rawson
>             Fix For: 0.20.1, 0.21.0
>
>         Attachments: HBASE-1795.patch
>
>
> during log recovery, we dont reset Store.maxSeqId, thus new log entries are stamped starting off from the old files.  This can cause a problem if we fail and recover again, since the new mutations are deemed "old" and shouldnt be applied in a subsequent recovery scenario.

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

[jira] Updated: (HBASE-1795) log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'

Posted by "ryan rawson (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-1795?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

ryan rawson updated HBASE-1795:
-------------------------------

    Attachment: HBASE-1795.patch

also fixes 1794

> log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'
> ------------------------------------------------------------------------------------------
>
>                 Key: HBASE-1795
>                 URL: https://issues.apache.org/jira/browse/HBASE-1795
>             Project: Hadoop HBase
>          Issue Type: Bug
>    Affects Versions: 0.20.0, 0.21.0
>            Reporter: ryan rawson
>             Fix For: 0.20.1, 0.21.0
>
>         Attachments: HBASE-1795.patch
>
>
> during log recovery, we dont reset Store.maxSeqId, thus new log entries are stamped starting off from the old files.  This can cause a problem if we fail and recover again, since the new mutations are deemed "old" and shouldnt be applied in a subsequent recovery scenario.

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

[jira] Commented: (HBASE-1795) log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'

Posted by "stack (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-1795?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12754293#action_12754293 ] 

stack commented on HBASE-1795:
------------------------------

+1 committing....

> log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'
> ------------------------------------------------------------------------------------------
>
>                 Key: HBASE-1795
>                 URL: https://issues.apache.org/jira/browse/HBASE-1795
>             Project: Hadoop HBase
>          Issue Type: Bug
>    Affects Versions: 0.20.0, 0.21.0
>            Reporter: ryan rawson
>             Fix For: 0.20.1, 0.21.0
>
>         Attachments: HBASE-1795.patch
>
>
> during log recovery, we dont reset Store.maxSeqId, thus new log entries are stamped starting off from the old files.  This can cause a problem if we fail and recover again, since the new mutations are deemed "old" and shouldnt be applied in a subsequent recovery scenario.

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

[jira] Commented: (HBASE-1795) log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'

Posted by "Jean-Daniel Cryans (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-1795?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12753899#action_12753899 ] 

Jean-Daniel Cryans commented on HBASE-1795:
-------------------------------------------

+1

> log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'
> ------------------------------------------------------------------------------------------
>
>                 Key: HBASE-1795
>                 URL: https://issues.apache.org/jira/browse/HBASE-1795
>             Project: Hadoop HBase
>          Issue Type: Bug
>    Affects Versions: 0.20.0, 0.21.0
>            Reporter: ryan rawson
>             Fix For: 0.20.1, 0.21.0
>
>         Attachments: HBASE-1795.patch
>
>
> during log recovery, we dont reset Store.maxSeqId, thus new log entries are stamped starting off from the old files.  This can cause a problem if we fail and recover again, since the new mutations are deemed "old" and shouldnt be applied in a subsequent recovery scenario.

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

[jira] Resolved: (HBASE-1795) log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'

Posted by "stack (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-1795?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

stack resolved HBASE-1795.
--------------------------

    Resolution: Fixed

Committed branch and trunk

> log recovery doesnt reset the max sequence id, new logfiles can get tossed as 'duplicates'
> ------------------------------------------------------------------------------------------
>
>                 Key: HBASE-1795
>                 URL: https://issues.apache.org/jira/browse/HBASE-1795
>             Project: Hadoop HBase
>          Issue Type: Bug
>    Affects Versions: 0.20.0, 0.21.0
>            Reporter: ryan rawson
>            Assignee: ryan rawson
>             Fix For: 0.20.1, 0.21.0
>
>         Attachments: HBASE-1795.patch
>
>
> during log recovery, we dont reset Store.maxSeqId, thus new log entries are stamped starting off from the old files.  This can cause a problem if we fail and recover again, since the new mutations are deemed "old" and shouldnt be applied in a subsequent recovery scenario.

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