You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "Jean-Daniel Cryans (JIRA)" <ji...@apache.org> on 2008/10/27 18:56:44 UTC
[jira] Created: (HBASE-963) Fix the retries in HTable.flushCommit
Fix the retries in HTable.flushCommit
-------------------------------------
Key: HBASE-963
URL: https://issues.apache.org/jira/browse/HBASE-963
Project: Hadoop HBase
Issue Type: Bug
Components: client
Reporter: Jean-Daniel Cryans
Assignee: Jean-Daniel Cryans
Priority: Blocker
Fix For: 0.19.0
Many are seeing writes not getting into tables since HBASE-748 was committed. As I explained to Jonathan:
<jdcryans> jgray: but I'm pretty sure that under high load a split took too long and the write didn't get through
<jdcryans> jgray: That part has to be refactored to be used with other kind of batching
<jdcryans> jgray: because the exponential backoff retry logic is in HCM but I had to get the retry out of it...
So this issue is about getting the retry logic out of HTable, hopefully merging it with the rest of HCM.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-963) Fix the retries in HTable.flushCommit
Posted by "Jean-Daniel Cryans (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-963?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jean-Daniel Cryans updated HBASE-963:
-------------------------------------
Attachment: hbase-963-ugly.patch
This patch shows how I plan to do the refactoring. Without HBASE-880, it's very ugly but it should fix the rows that are not getting committed.
> Fix the retries in HTable.flushCommit
> -------------------------------------
>
> Key: HBASE-963
> URL: https://issues.apache.org/jira/browse/HBASE-963
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Reporter: Jean-Daniel Cryans
> Assignee: Jean-Daniel Cryans
> Priority: Blocker
> Fix For: 0.19.0
>
> Attachments: hbase-963-ugly.patch
>
>
> Many are seeing writes not getting into tables since HBASE-748 was committed. As I explained to Jonathan:
> <jdcryans> jgray: but I'm pretty sure that under high load a split took too long and the write didn't get through
> <jdcryans> jgray: That part has to be refactored to be used with other kind of batching
> <jdcryans> jgray: because the exponential backoff retry logic is in HCM but I had to get the retry out of it...
> So this issue is about getting the retry logic out of HTable, hopefully merging it with the rest of HCM.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (HBASE-963) Fix the retries in HTable.flushCommit
Posted by "Jean-Daniel Cryans (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-963?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jean-Daniel Cryans resolved HBASE-963.
--------------------------------------
Resolution: Fixed
Committed to trunk. Thanks for testing Jonathan!
> Fix the retries in HTable.flushCommit
> -------------------------------------
>
> Key: HBASE-963
> URL: https://issues.apache.org/jira/browse/HBASE-963
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Reporter: Jean-Daniel Cryans
> Assignee: Jean-Daniel Cryans
> Priority: Blocker
> Fix For: 0.19.0
>
> Attachments: hbase-963-ugly.patch, hbase-963-v2.patch
>
>
> Many are seeing writes not getting into tables since HBASE-748 was committed. As I explained to Jonathan:
> <jdcryans> jgray: but I'm pretty sure that under high load a split took too long and the write didn't get through
> <jdcryans> jgray: That part has to be refactored to be used with other kind of batching
> <jdcryans> jgray: because the exponential backoff retry logic is in HCM but I had to get the retry out of it...
> So this issue is about getting the retry logic out of HTable, hopefully merging it with the rest of HCM.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-963) Fix the retries in HTable.flushCommit
Posted by "Jean-Daniel Cryans (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-963?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jean-Daniel Cryans updated HBASE-963:
-------------------------------------
Attachment: hbase-963-v2.patch
Updated after HBASE-966. Passes PE, accounts for all rows when auto flush is enabled and when it's not.
> Fix the retries in HTable.flushCommit
> -------------------------------------
>
> Key: HBASE-963
> URL: https://issues.apache.org/jira/browse/HBASE-963
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Reporter: Jean-Daniel Cryans
> Assignee: Jean-Daniel Cryans
> Priority: Blocker
> Fix For: 0.19.0
>
> Attachments: hbase-963-ugly.patch, hbase-963-v2.patch
>
>
> Many are seeing writes not getting into tables since HBASE-748 was committed. As I explained to Jonathan:
> <jdcryans> jgray: but I'm pretty sure that under high load a split took too long and the write didn't get through
> <jdcryans> jgray: That part has to be refactored to be used with other kind of batching
> <jdcryans> jgray: because the exponential backoff retry logic is in HCM but I had to get the retry out of it...
> So this issue is about getting the retry logic out of HTable, hopefully merging it with the rest of HCM.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-963) Fix the retries in HTable.flushCommit
Posted by "Jonathan Gray (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-963?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12644128#action_12644128 ]
Jonathan Gray commented on HBASE-963:
-------------------------------------
+1 no issues for 24 hours
> Fix the retries in HTable.flushCommit
> -------------------------------------
>
> Key: HBASE-963
> URL: https://issues.apache.org/jira/browse/HBASE-963
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Reporter: Jean-Daniel Cryans
> Assignee: Jean-Daniel Cryans
> Priority: Blocker
> Fix For: 0.19.0
>
> Attachments: hbase-963-ugly.patch, hbase-963-v2.patch
>
>
> Many are seeing writes not getting into tables since HBASE-748 was committed. As I explained to Jonathan:
> <jdcryans> jgray: but I'm pretty sure that under high load a split took too long and the write didn't get through
> <jdcryans> jgray: That part has to be refactored to be used with other kind of batching
> <jdcryans> jgray: because the exponential backoff retry logic is in HCM but I had to get the retry out of it...
> So this issue is about getting the retry logic out of HTable, hopefully merging it with the rest of HCM.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-963) Fix the retries in HTable.flushCommit
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-963?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12643015#action_12643015 ]
stack commented on HBASE-963:
-----------------------------
Patch is fine -- would suggest that before commit, change the close so it just notes exception rather than close it since on close, there is nothing much that can be done about and IOE.
> Fix the retries in HTable.flushCommit
> -------------------------------------
>
> Key: HBASE-963
> URL: https://issues.apache.org/jira/browse/HBASE-963
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Reporter: Jean-Daniel Cryans
> Assignee: Jean-Daniel Cryans
> Priority: Blocker
> Fix For: 0.19.0
>
> Attachments: hbase-963-ugly.patch
>
>
> Many are seeing writes not getting into tables since HBASE-748 was committed. As I explained to Jonathan:
> <jdcryans> jgray: but I'm pretty sure that under high load a split took too long and the write didn't get through
> <jdcryans> jgray: That part has to be refactored to be used with other kind of batching
> <jdcryans> jgray: because the exponential backoff retry logic is in HCM but I had to get the retry out of it...
> So this issue is about getting the retry logic out of HTable, hopefully merging it with the rest of HCM.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.