[jira] [Commented] (CASSANDRA-5395) Compaction doesn't remove index entries as designed

2013-03-29 Thread Jonathan Ellis (JIRA)

[ 
https://issues.apache.org/jira/browse/CASSANDRA-5395?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13617522#comment-13617522
 ] 

Jonathan Ellis commented on CASSANDRA-5395:
---

Committed with suggested improvements.

 Compaction doesn't remove index entries as designed
 ---

 Key: CASSANDRA-5395
 URL: https://issues.apache.org/jira/browse/CASSANDRA-5395
 Project: Cassandra
  Issue Type: Bug
  Components: Core
Affects Versions: 1.2.0
Reporter: Jonathan Ellis
Assignee: Jonathan Ellis
Priority: Minor
 Fix For: 1.2.4

 Attachments: 5395-2.txt, 5395.txt


 PerColumnIndexUpdater ignores updates where the new value is a tombstone.  It 
 should still remove the index entry on oldColumn.
 (Note that this will not affect user-visible correctness, since 
 KeysSearcher/CompositeSearcher will issue deletes against stale index 
 entries, but having more stale entries than we should could affect 
 performance.)

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


[jira] [Commented] (CASSANDRA-5395) Compaction doesn't remove index entries as designed

2013-03-28 Thread Sam Tunnicliffe (JIRA)

[ 
https://issues.apache.org/jira/browse/CASSANDRA-5395?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13616605#comment-13616605
 ] 

Sam Tunnicliffe commented on CASSANDRA-5395:


lgtm, just have 2 trivial queries:

In LCR  PCR, if the purpose of the additional clauses is to omit unnecessary 
column lookups, should the column lookup be the last of the 'd  conditions?

{code}
 if (indexer != SecondaryIndexManager.nullUpdater
 !column.isMarkedForDelete()
 container.getColumn(column.name()) != column)
{code}


Class documentation in IdentityQueryFilter states Only for use in testing; 
will read entire CF into memory. Seeing as its being used in non-test code we 
should probably amend the docstring


 Compaction doesn't remove index entries as designed
 ---

 Key: CASSANDRA-5395
 URL: https://issues.apache.org/jira/browse/CASSANDRA-5395
 Project: Cassandra
  Issue Type: Bug
  Components: Core
Affects Versions: 1.2.0
Reporter: Jonathan Ellis
Priority: Minor
 Fix For: 1.2.4

 Attachments: 5395-2.txt, 5395.txt


 PerColumnIndexUpdater ignores updates where the new value is a tombstone.  It 
 should still remove the index entry on oldColumn.
 (Note that this will not affect user-visible correctness, since 
 KeysSearcher/CompositeSearcher will issue deletes against stale index 
 entries, but having more stale entries than we should could affect 
 performance.)

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


[jira] [Commented] (CASSANDRA-5395) Compaction doesn't remove index entries as designed

2013-03-27 Thread Jonathan Ellis (JIRA)

[ 
https://issues.apache.org/jira/browse/CASSANDRA-5395?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13615816#comment-13615816
 ] 

Jonathan Ellis commented on CASSANDRA-5395:
---

Fixed a test failure and pushed to 
http://github.com/jbellis/cassandra/tree/5395.  (Original patches were also 
half 1.2, half trunk.  Now both against 1.2.)

 Compaction doesn't remove index entries as designed
 ---

 Key: CASSANDRA-5395
 URL: https://issues.apache.org/jira/browse/CASSANDRA-5395
 Project: Cassandra
  Issue Type: Bug
  Components: Core
Affects Versions: 1.2.0
Reporter: Jonathan Ellis
Priority: Minor
 Fix For: 1.2.4

 Attachments: 5395-2.txt, 5395.txt


 PerColumnIndexUpdater ignores updates where the new value is a tombstone.  It 
 should still remove the index entry on oldColumn.
 (Note that this will not affect user-visible correctness, since 
 KeysSearcher/CompositeSearcher will issue deletes against stale index 
 entries, but having more stale entries than we should could affect 
 performance.)

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira


[jira] [Commented] (CASSANDRA-5395) Compaction doesn't remove index entries as designed

2013-03-27 Thread Jonathan Ellis (JIRA)

[ 
https://issues.apache.org/jira/browse/CASSANDRA-5395?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13615972#comment-13615972
 ] 

Jonathan Ellis commented on CASSANDRA-5395:
---

... And pushed a 3rd commit applying the fix in patch to to 
ParallelCompactionIterable.

This also has the side effect of switching from TSM to ABSM for PrecompactedRow 
and ParallelCompactionIterable, so there may be a performance improvement.

 Compaction doesn't remove index entries as designed
 ---

 Key: CASSANDRA-5395
 URL: https://issues.apache.org/jira/browse/CASSANDRA-5395
 Project: Cassandra
  Issue Type: Bug
  Components: Core
Affects Versions: 1.2.0
Reporter: Jonathan Ellis
Priority: Minor
 Fix For: 1.2.4

 Attachments: 5395-2.txt, 5395.txt


 PerColumnIndexUpdater ignores updates where the new value is a tombstone.  It 
 should still remove the index entry on oldColumn.
 (Note that this will not affect user-visible correctness, since 
 KeysSearcher/CompositeSearcher will issue deletes against stale index 
 entries, but having more stale entries than we should could affect 
 performance.)

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira