[jira] [Commented] (CASSANDRA-16968) Diff Job retry bug fixes in reading previous run's job parameters & marking the job status

["Jyothsna Konisa (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-16968?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17416948#comment-17416948 ] 

Jyothsna Konisa commented on CASSANDRA-16968:
---------------------------------------------

[https://github.com/apache/cassandra-diff/pull/16]
 # The try resource block in DiffJob.java closes the session object before the code in exception or finally block gets executed. We are marking the job as not running in the exception block which throws an exception as the session object is already closed. Changing the resource try catch block to try catch finally block so that session object will not be closed until cleanup is complete.
 # When job_id is passed as a config property for the first time, we will not have metadata associated with job_id in metadata table but the current code attempts to get the job metadata for the passed jobId and as those details will not be present, a null pointer exception is thrown. This patch fixes this issue by getting jobParameters from the table only when they are available otherwise creates new job parameters with passed JobId or random UUID.

> Diff Job retry bug fixes in reading previous run's job parameters & marking the job status
> ------------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-16968
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-16968
>             Project: Cassandra
>          Issue Type: New Feature
>            Reporter: Jyothsna Konisa
>            Assignee: Jyothsna Konisa
>            Priority: Normal
>
> Diff job retry with same jobId should avoid running diffs on the partitions that were successfully diffed previously. The retry failed because previous run failed to mark the job as not running on exit. This is due to a bug in the resource try catch block where session object Is closed before marking the job as not running. Also there is another bug in the way we get job parameters during rerun of a failed diff job.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org