You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Amitanand Aiyer (Created) (JIRA)" <ji...@apache.org> on 2012/02/16 23:17:00 UTC
[jira] [Created] (HBASE-5418) use different memstoreTS for
different operations in the same RowMutation.
use different memstoreTS for different operations in the same RowMutation.
--------------------------------------------------------------------------
Key: HBASE-5418
URL: https://issues.apache.org/jira/browse/HBASE-5418
Project: HBase
Issue Type: Sub-task
Reporter: Amitanand Aiyer
Assignee: Amitanand Aiyer
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (HBASE-5418) use different memstoreTS for
different operations in the same RowMutation.
Posted by "Amitanand Aiyer (Updated) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-5418?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amitanand Aiyer updated HBASE-5418:
-----------------------------------
Description:
Assigning different memstoreTS will enable us to guarantee that the
operations will appear to take effect, along the same order, in which
they were added to create the RowMutation.
Based on the diff after renaming to RowMutations.
> use different memstoreTS for different operations in the same RowMutation.
> --------------------------------------------------------------------------
>
> Key: HBASE-5418
> URL: https://issues.apache.org/jira/browse/HBASE-5418
> Project: HBase
> Issue Type: Sub-task
> Components: client, coprocessors, regionserver
> Reporter: Amitanand Aiyer
> Assignee: Amitanand Aiyer
> Fix For: 0.94.0
>
>
> Assigning different memstoreTS will enable us to guarantee that the
> operations will appear to take effect, along the same order, in which
> they were added to create the RowMutation.
> Based on the diff after renaming to RowMutations.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-5418) use different memstoreTS for
different operations in the same RowMutation.
Posted by "Lars Hofhansl (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-5418?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13210072#comment-13210072 ]
Lars Hofhansl commented on HBASE-5418:
--------------------------------------
This would address some (but not all) of my concerns with HBASE-5241.
I still find the notion that we have an order that is different from the one implied by the timestamps "dangerous", especially because the sort order of the KVs is not according to the memstoreTSs.
> use different memstoreTS for different operations in the same RowMutation.
> --------------------------------------------------------------------------
>
> Key: HBASE-5418
> URL: https://issues.apache.org/jira/browse/HBASE-5418
> Project: HBase
> Issue Type: Sub-task
> Components: client, coprocessors, regionserver
> Reporter: Amitanand Aiyer
> Assignee: Amitanand Aiyer
> Fix For: 0.94.0
>
> Attachments: HBASE-5418.D1761.1.patch
>
>
> Assigning different memstoreTS will enable us to guarantee that the
> operations will appear to take effect, along the same order, in which
> they were added to create the RowMutation.
> Based on the diff after renaming to RowMutations.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-5418) use different memstoreTS for
different operations in the same RowMutation.
Posted by "Phabricator (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-5418?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13219655#comment-13219655 ]
Phabricator commented on HBASE-5418:
------------------------------------
aaiyer has commented on the revision "HBASE-5418 [jira] use different memstoreTS for different operations in the same RowMutation.".
INLINE COMMENTS
src/main/java/org/apache/hadoop/hbase/regionserver/MultiVersionConsistencyControl.java:121 I'll do that.
REVISION DETAIL
https://reviews.facebook.net/D1761
> use different memstoreTS for different operations in the same RowMutation.
> --------------------------------------------------------------------------
>
> Key: HBASE-5418
> URL: https://issues.apache.org/jira/browse/HBASE-5418
> Project: HBase
> Issue Type: Sub-task
> Components: client, coprocessors, regionserver
> Reporter: Amitanand Aiyer
> Assignee: Amitanand Aiyer
> Fix For: 0.94.0
>
> Attachments: HBASE-5418.D1761.1.patch
>
>
> Assigning different memstoreTS will enable us to guarantee that the
> operations will appear to take effect, along the same order, in which
> they were added to create the RowMutation.
> Based on the diff after renaming to RowMutations.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (HBASE-5418) use different memstoreTS for
different operations in the same RowMutation.
Posted by "Phabricator (Updated) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-5418?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Phabricator updated HBASE-5418:
-------------------------------
Attachment: HBASE-5418.D1761.1.patch
aaiyer requested code review of "HBASE-5418 [jira] use different memstoreTS for different operations in the same RowMutation.".
Reviewers: JIRA
https://issues.apache.org/jira/browse/HBASE-5418
Use different memstoreTS for distinct operations in the same RowMutation.
Assigning different memstoreTS will enable us to guarantee that the
operations will appear to take effect, along the same order, in which
they were added to create the RowMutation.
Based on the diff after renaming to RowMutations.
TEST PLAN
EMPTY
REVISION DETAIL
https://reviews.facebook.net/D1761
AFFECTED FILES
src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.java
src/main/java/org/apache/hadoop/hbase/regionserver/MultiVersionConsistencyControl.java
MANAGE HERALD DIFFERENTIAL RULES
https://reviews.facebook.net/herald/view/differential/
WHY DID I GET THIS EMAIL?
https://reviews.facebook.net/herald/transcript/3753/
Tip: use the X-Herald-Rules header to filter Herald messages in your client.
> use different memstoreTS for different operations in the same RowMutation.
> --------------------------------------------------------------------------
>
> Key: HBASE-5418
> URL: https://issues.apache.org/jira/browse/HBASE-5418
> Project: HBase
> Issue Type: Sub-task
> Components: client, coprocessors, regionserver
> Reporter: Amitanand Aiyer
> Assignee: Amitanand Aiyer
> Fix For: 0.94.0
>
> Attachments: HBASE-5418.D1761.1.patch
>
>
> Assigning different memstoreTS will enable us to guarantee that the
> operations will appear to take effect, along the same order, in which
> they were added to create the RowMutation.
> Based on the diff after renaming to RowMutations.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (HBASE-5418) use different memstoreTS for
different operations in the same RowMutation.
Posted by "Lars Hofhansl (Updated) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-5418?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Lars Hofhansl updated HBASE-5418:
---------------------------------
Fix Version/s: (was: 0.94.0)
0.96.0
Moving out of 0.94.
> use different memstoreTS for different operations in the same RowMutation.
> --------------------------------------------------------------------------
>
> Key: HBASE-5418
> URL: https://issues.apache.org/jira/browse/HBASE-5418
> Project: HBase
> Issue Type: Sub-task
> Components: client, coprocessors, regionserver
> Reporter: Amitanand Aiyer
> Assignee: Amitanand Aiyer
> Fix For: 0.96.0
>
> Attachments: HBASE-5418.D1761.1.patch
>
>
> Assigning different memstoreTS will enable us to guarantee that the
> operations will appear to take effect, along the same order, in which
> they were added to create the RowMutation.
> Based on the diff after renaming to RowMutations.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-5418) use different memstoreTS for
different operations in the same RowMutation.
Posted by "Phabricator (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-5418?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13209906#comment-13209906 ]
Phabricator commented on HBASE-5418:
------------------------------------
tedyu has commented on the revision "HBASE-5418 [jira] use different memstoreTS for different operations in the same RowMutation.".
INLINE COMMENTS
src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.java:4245 Should ctr be lifted to above the try block so that we only call mvcc.completeMemstoreInsert() for the WriteEntry's we actually touched ?
src/main/java/org/apache/hadoop/hbase/regionserver/MultiVersionConsistencyControl.java:121 Should we simplify beginMemstoreInsert() by letting it call this new method ?
This would allow maintaining one method instead of two in case we add something in the future.
REVISION DETAIL
https://reviews.facebook.net/D1761
> use different memstoreTS for different operations in the same RowMutation.
> --------------------------------------------------------------------------
>
> Key: HBASE-5418
> URL: https://issues.apache.org/jira/browse/HBASE-5418
> Project: HBase
> Issue Type: Sub-task
> Components: client, coprocessors, regionserver
> Reporter: Amitanand Aiyer
> Assignee: Amitanand Aiyer
> Fix For: 0.94.0
>
> Attachments: HBASE-5418.D1761.1.patch
>
>
> Assigning different memstoreTS will enable us to guarantee that the
> operations will appear to take effect, along the same order, in which
> they were added to create the RowMutation.
> Based on the diff after renaming to RowMutations.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HBASE-5418) use different memstoreTS for
different operations in the same RowMutation.
Posted by "Phabricator (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-5418?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13219654#comment-13219654 ]
Phabricator commented on HBASE-5418:
------------------------------------
aaiyer has commented on the revision "HBASE-5418 [jira] use different memstoreTS for different operations in the same RowMutation.".
INLINE COMMENTS
src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.java:4245 I think we should completeMemstoreInsert for all the writeEntries, regardless of weather they suceed or not.
If we do not call completeMemstoreInsert (for a incomplete operation), then none of the puts/deletes that start after that operation can ever complete.
As long as we are "totally giving up" on retrying the operation, technically, it should be fine to call completeMemstoreInsert. All it does, as far as I can tell, is to remove the writeEntry from the list of "current writes".
(Although, from code inspection, it seems unlikely why one will suceed and the other does not --- because applyFamilyMapToMemstore does not seem to throw exceptions).
REVISION DETAIL
https://reviews.facebook.net/D1761
> use different memstoreTS for different operations in the same RowMutation.
> --------------------------------------------------------------------------
>
> Key: HBASE-5418
> URL: https://issues.apache.org/jira/browse/HBASE-5418
> Project: HBase
> Issue Type: Sub-task
> Components: client, coprocessors, regionserver
> Reporter: Amitanand Aiyer
> Assignee: Amitanand Aiyer
> Fix For: 0.94.0
>
> Attachments: HBASE-5418.D1761.1.patch
>
>
> Assigning different memstoreTS will enable us to guarantee that the
> operations will appear to take effect, along the same order, in which
> they were added to create the RowMutation.
> Based on the diff after renaming to RowMutations.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira