[
https://issues.apache.org/jira/browse/CASSANDRA-15807?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17108672#comment-17108672
]
Yifan Cai commented on CASSANDRA-15807:
---------------------------------------
PR: https://github.com/apache/cassandra-diff/pull/8
Code: https://github.com/yifan-c/cassandra-diff/tree/support-multiple-keyspaces
At a high level, in order to support comparing tables among multiple keyspaces,
it now uses <keyspace, table> pair to locate the table. The compare logic
remains the same.
The changes is relatively invasive as it changes the input format and the
schema for job metadata. Given the tool is at its early stage, I think it is
OK. Feel free to raise it if any objections.
End-to-end testing performed locally to verify the spark job runs and api
service is working as expected.
Changed items:
* Replaced discrete inputs, keyspace and tables, with one list of keyspace
table pairs. Now a table is identified by keyspace and table name.
* The scheme of JobMetadataDb is updated correspondingly to replace all
'table_name' with 'keyspace_table_name' and replace table_names with
keyspace_table_names
* DBService in api service is changed to make sure it can query with the new
schema.
* Updated the ReadMe with an example of running with multiple keyspaces.
> Support multiple keyspaces in Cassandra-Diff
> --------------------------------------------
>
> Key: CASSANDRA-15807
> URL: https://issues.apache.org/jira/browse/CASSANDRA-15807
> Project: Cassandra
> Issue Type: Improvement
> Components: Tool/diff
> Reporter: Yifan Cai
> Assignee: Yifan Cai
> Priority: Normal
>
> Adding the support to run diff comparison of tables across multiple keyspaces
> to avoid bringing up multiple spark jobs and give a better view of data diffs
> in the whole cluster.
> Cassandra-diff currently only support compare multiple tables under one
> single keyspace.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
