You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "AJ Slater (JIRA)" <ji...@apache.org> on 2010/06/01 01:25:37 UTC
[jira] Created: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Reading with CL > ONE returns multiple copies of the same column per key.
-------------------------------------------------------------------------
Key: CASSANDRA-1145
URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
Project: Cassandra
Issue Type: Bug
Components: Core
Affects Versions: 0.6.1
Environment: ubuntu jaunty
Reporter: AJ Slater
Testing with 0.6-trunk today:
Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
This is with pycassa get_range() which is using get_range_slice()
I see the same behavior with 0.6.1 and 0.6.2 debs
If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jeremy Hanna (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873832#action_12873832 ]
Jeremy Hanna commented on CASSANDRA-1145:
-----------------------------------------
Okay yeah - that fix was in revision 948934, so it looks like it's a different issue. Tx.
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "AJ Slater (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873829#action_12873829 ]
AJ Slater commented on CASSANDRA-1145:
--------------------------------------
I saw the same behavior with cassandra-0.6 svn revision 949892
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jeremy Hanna (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873827#action_12873827 ]
Jeremy Hanna commented on CASSANDRA-1145:
-----------------------------------------
Hi AJ, could you try to replicate it with the latest 0.6 branch just to see if it wasn't fixed by another fix (cassandra-1042)? Just wondering if it's a the same bug but showing itself in more than one place.
http://svn.apache.org/repos/asf/cassandra/branches/cassandra-0.6/
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Folke Behrens (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Folke Behrens updated CASSANDRA-1145:
-------------------------------------
Attachment: demo.patch
CollatingIterator expects both collections to be sorted. See demo.patch.
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.5
>
> Attachments: 0001-Added-a-unit-test.patch, bugtest.py, demo.patch, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jeremy Hanna (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jeremy Hanna reassigned CASSANDRA-1145:
---------------------------------------
Assignee: Jeremy Hanna
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.5
>
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Ellis reassigned CASSANDRA-1145:
-----------------------------------------
Assignee: Jeremy Hanna
Joost: :) no, it really is a bug
Jeremy: could you have a look?
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.3
>
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Ellis updated CASSANDRA-1145:
--------------------------------------
Fix Version/s: 0.6.5
(was: 0.6.4)
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.5
>
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Ellis updated CASSANDRA-1145:
--------------------------------------
Attachment: 1145-v2.txt
The order we read them in CFS [token order] is the "right" order. Stomping on that will take as back to CASSANDRA-1042 type bugs.
Does v2 (attached) fix the problem?
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.5
>
> Attachments: 0001-Added-a-unit-test.patch, 0001-Sort-keys.patch, 1145-v2.txt, bugtest.py, demo.patch, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Joost Ouwerkerk (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12879505#action_12879505 ]
Joost Ouwerkerk commented on CASSANDRA-1145:
--------------------------------------------
We're getting duplicate rows from get_range_slice when NOT using ConsistencyLevel.ONE. That is to say, when using QUORUM or ALL, we're getting duplicate rows, but when using ONE, we are not. Is this case misnamed?
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jeremy Hanna (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12885744#action_12885744 ]
Jeremy Hanna commented on CASSANDRA-1145:
-----------------------------------------
Hi AJ, taking another look at this. I'm getting errors when I try to put ('localhost:9160') in the SERVERS variable in bugtest.py. Specifically, I'm seeing that pycassa is giving a NoServerAvailable exception when trying to connect. So I'm unable to reproduce the problem.
Can you see if you can get it to run and hopefully replicate the problem just with the localhost? I'm not sure if I'm just configuring that incorrectly either.
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.4
>
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jeremy Hanna (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12885786#action_12885786 ]
Jeremy Hanna commented on CASSANDRA-1145:
-----------------------------------------
Sweet sassafras - thanks AJ, that was it - matching the localhost in the SERVERS variable to localhost in the rpc address in storage-conf.xml. Looking at it more now.
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.4
>
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jeremy Hanna (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jeremy Hanna updated CASSANDRA-1145:
------------------------------------
Attachment: 0001-Added-a-unit-test.patch
Added a unit test that will expose the problem.
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.5
>
> Attachments: 0001-Added-a-unit-test.patch, bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Ellis updated CASSANDRA-1145:
--------------------------------------
Fix Version/s: 0.6.3
Priority: Critical (was: Major)
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Priority: Critical
> Fix For: 0.6.3
>
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jeremy Hanna (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12895449#action_12895449 ]
Jeremy Hanna commented on CASSANDRA-1145:
-----------------------------------------
Folke - you're right. I'll focus on the creation of the list of rows. I had thought we looked at the ordering, but I think since they were in the same order, we didn't think twice. Good catch.
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.5
>
> Attachments: 0001-Added-a-unit-test.patch, bugtest.py, demo.patch, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jeremy Hanna (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12895678#action_12895678 ]
Jeremy Hanna commented on CASSANDRA-1145:
-----------------------------------------
Yes - I wondered if what I had done would affect other things, such as reverting other changes. Thanks. I could see where the problem was, but wasn't 100% sure where to make the change. Sounds like they were sorted correctly wrt to token, which our CollatingIterator needed to take into account.
I ran the same tests again that could formerly reproduce the problem (with several permutations) and it fixes the problem.
I also ran `ant test` and `nosetests` to make sure everything else was happy.
+1 from me.
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.5
>
> Attachments: 0001-Added-a-unit-test.patch, 0001-Sort-keys.patch, 1145-v2.txt, bugtest.py, demo.patch, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jeremy Hanna (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jeremy Hanna updated CASSANDRA-1145:
------------------------------------
Attachment: 0001-Sort-keys.patch
Adding a fix patch that will sort the keys before getRangeSlice uses them to get their associated rows. That will keep them sorted when they are sent back to the client, for the CollatingIterator.
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.5
>
> Attachments: 0001-Added-a-unit-test.patch, 0001-Sort-keys.patch, bugtest.py, demo.patch, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Ellis reassigned CASSANDRA-1145:
-----------------------------------------
Assignee: (was: Jeremy Hanna)
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Priority: Critical
> Fix For: 0.6.5
>
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "AJ Slater (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12885747#action_12885747 ]
AJ Slater commented on CASSANDRA-1145:
--------------------------------------
Looks like pycassa has a problem with tuples with one element, which is another problem. So to fix, use a list:
SERVERS= ['localhost:9160']
Also I ran into your NoServerAvailable problem, using localhost:9160.
Cassandra is likely not really listening on 'localhost:9160', try 'myServerName:9160' or use the exact IP you specify to thrift in storage-conf.xml
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Priority: Critical
> Fix For: 0.6.4
>
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "AJ Slater (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
AJ Slater updated CASSANDRA-1145:
---------------------------------
Attachment: bugtest.py
storage-conf.xml
attached is a storage-conf and example python program.
it requires the python thrift bindings and pycassa: http://github.com/vomjom/pycassa
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Issue Comment Edited: (CASSANDRA-1145) Reading with CL > ONE
returns multiple copies of the same column per key.
Posted by "AJ Slater (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12873823#action_12873823 ]
AJ Slater edited comment on CASSANDRA-1145 at 5/31/10 7:32 PM:
---------------------------------------------------------------
attached is a storage-conf and example python program.
it requires the python thrift bindings and pycassa: http://github.com/vomjom/pycassa
usage: ./bugtest.py [TEXT1 TEXT2..]
arguments on the command line are inserted into cassandra and then range_get is used to get and pretty print them, with timestamps.
it counts the multiple copies of the same rows at the end.
was (Author: ajslater):
attached is a storage-conf and example python program.
it requires the python thrift bindings and pycassa: http://github.com/vomjom/pycassa
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6.1
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Attachments: bugtest.py, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (CASSANDRA-1145) Reading with CL > ONE returns
multiple copies of the same column per key.
Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-1145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Ellis updated CASSANDRA-1145:
--------------------------------------
Affects Version/s: 0.6
(was: 0.6.1)
Priority: Major (was: Critical)
> Reading with CL > ONE returns multiple copies of the same column per key.
> -------------------------------------------------------------------------
>
> Key: CASSANDRA-1145
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1145
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6
> Environment: ubuntu jaunty
> Reporter: AJ Slater
> Assignee: Jeremy Hanna
> Fix For: 0.6.5
>
> Attachments: 0001-Added-a-unit-test.patch, 0001-Sort-keys.patch, 1145-v2.txt, bugtest.py, demo.patch, storage-conf.xml
>
>
> Testing with 0.6-trunk today:
> Reading with CL > ONE returns multiple copies of the same column per key consistent with the replicas queried before return. i.e, for RC=3, a QUORUM read yields 2 copies and an ALL read returns 3.
> This is with pycassa get_range() which is using get_range_slice()
> I see the same behavior with 0.6.1 and 0.6.2 debs
> If my experience is not unique, anyone using get_range_slice is now deluged with duplicate data.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.