You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Tyler Hobbs (JIRA)" <ji...@apache.org> on 2015/07/17 21:49:06 UTC
[jira] [Comment Edited] (CASSANDRA-9704) On-wire backward
compatibility for 8099
[ https://issues.apache.org/jira/browse/CASSANDRA-9704?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14631805#comment-14631805 ]
Tyler Hobbs edited comment on CASSANDRA-9704 at 7/17/15 7:48 PM:
-----------------------------------------------------------------
There are two remaining failing tests (that I'm aware of) in the new {{upgrade_tests/cql_tests.py}}:
* {{only_pk_test}}: this requires a schema migration fix in CASSANDRA-6717. This [this comment|https://issues.apache.org/jira/browse/CASSANDRA-6717?focusedCommentId=14626965&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14626965] for details. I've added a {{@require()}} skip to this test for now.
* {{noncomposite_static_cf_test}}: For the query {{"SELECT * FROM users WHERE userid = f47ac10b-58cc-4372-a567-0e02b2c3d479"}}, the 2.1 node sends a read command that gets translated to this in 3.0: {{Read(ks.users columns=age, firstname, lastname rowFilter= limits= key=550e8400-e29b-41d4-a716-446655440000 filter=names(), nowInSec=1437159837)}}. Normally in 3.0 we would use a slice query for this, but the above query _should_ work. However, there appears to be some sort of internal problem handling the query, resulting in the response containing three cells: firstname, lastname, and a duplicate of lastname. I'm guessing this is caused by some misuse of a Reusable object, but it's damned hard to track down. I suspect that CASSANDRA-9705 would fix this, but I haven't verified this yet (see comments below about merging), so I've added a comment and left the test failing.
On the topic of CASSANDRA-9705, it looks like this branch is going to require quite a few changes to accommodate that. I suggest waiting until 9705 has been committed before rebasing and committing this branch.
was (Author: thobbs):
There are two remaining failing tests (that I'm aware of) in the new {{upgrade_tests/cql_tests.py}}:
* {{only_pk_test}}: this requires a schema migration fix in CASSANDRA-6717. This [this comment|https://issues.apache.org/jira/browse/CASSANDRA-6717?focusedCommentId=14626965&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14626965] for details. I've added a {{@require()}} skip to this test for now.
* {{noncomposite_static_cf_test}}: For the query {{"SELECT * FROM users WHERE userid = f47ac10b-58cc-4372-a567-0e02b2c3d479"}}, the 2.1 node sends a read command that gets translated to this in 3.0: {{Read(ks.users columns=age, firstname, lastname rowFilter= limits= key=550e8400-e29b-41d4-a716-446655440000 filter=names(), nowInSec=1437159837). Normally in 3.0 we would use a slice query for this, but the above query _should_ work. However, there appears to be some sort of internal problem handling the query, resulting in the response containing three cells: firstname, lastname, and a duplicate of lastname. I'm guessing this is caused by some misuse of a Reusable object, but it's damned hard to track down. I suspect that CASSANDRA-9705 would fix this, but I haven't verified this yet (see comments below about merging), so I've added a comment and left the test failing.
On the topic of CASSANDRA-9705, it looks like this branch is going to require quite a few changes to accommodate that. I suggest waiting until 9705 has been committed before rebasing and committing this branch.
> On-wire backward compatibility for 8099
> ---------------------------------------
>
> Key: CASSANDRA-9704
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9704
> Project: Cassandra
> Issue Type: Sub-task
> Reporter: Sylvain Lebresne
> Assignee: Tyler Hobbs
> Fix For: 3.0 beta 1
>
> Attachments: 9704-2.1.txt
>
>
> The currently committed patch for CASSANDRA-8099 has left backward compatibility on the wire as a TODO. This ticket is to track the actual doing (of which I know [~thobbs] has already done a good chunk).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)