You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Alex Petrov (JIRA)" <ji...@apache.org> on 2016/12/19 16:55:58 UTC

[jira] [Commented] (CASSANDRA-12373) 3.0 breaks CQL compatibility with super columns families

    [ https://issues.apache.org/jira/browse/CASSANDRA-12373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15761691#comment-15761691 ] 

Alex Petrov commented on CASSANDRA-12373:
-----------------------------------------

Previous patch had problems with renames of dense supercf (was a result of my assumption that renames would work 
similar to how they'd normally work in 3.x, although it's possible to rename all 4 columns in dense supercf and pk parts
 in sparse). Thanks [~iamaleksey] for clarifying sparse/dense semantics in the context of super-cf. 

Without renames, it was quite simple to sort out the supercolumn families, since the columns were always purely virtual.
With renames, it got a bit more difficult and several problems appeared:

  * difference between sparse and dense causes trouble during upgrade (since in once case we have multiple
regular rows, so we can't add the supercf key to regulars, since we won't be able to differentiate it later)
  * denseness calculation for supercf (because of the empty name map column) was causing errors during upgrade

There were several other smaller things, all covered in the patch.

|[3.X|https://github.com/ifesdjeen/cassandra/tree/12373-3.X]|[dtest|https://cassci.datastax.com/view/Dev/view/ifesdjeen/job/ifesdjeen-12373-3.X-dtest/]|[utest|https://cassci.datastax.com/view/Dev/view/ifesdjeen/job/ifesdjeen-12373-3.X-testall/]|
|[2.2|https://github.com/ifesdjeen/cassandra/tree/12373-2.2]|[dtest|https://cassci.datastax.com/view/Dev/view/ifesdjeen/job/ifesdjeen-12373-2.2-dtest/]|[utest|https://cassci.datastax.com/view/Dev/view/ifesdjeen/job/ifesdjeen-12373-2.2-testall/]|
|[dtest patch|https://github.com/ifesdjeen/cassandra-dtest/tree/12373]|

I didn't update trunk because of the [4881d9c308ccd6b5ca70925bf6ebedb70e7705fc|https://github.com/apache/cassandra/commit/4881d9c308ccd6b5ca70925bf6ebedb70e7705fc]

> 3.0 breaks CQL compatibility with super columns families
> --------------------------------------------------------
>
>                 Key: CASSANDRA-12373
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-12373
>             Project: Cassandra
>          Issue Type: Bug
>          Components: CQL
>            Reporter: Sylvain Lebresne
>            Assignee: Alex Petrov
>             Fix For: 3.0.x, 3.x
>
>
> This is a follow-up to CASSANDRA-12335 to fix the CQL side of super column compatibility.
> The details and a proposed solution can be found in the comments of CASSANDRA-12335 but the crux of the issue is that super column famillies show up differently in CQL in 3.0.x/3.x compared to 2.x, hence breaking backward compatibilty.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)