You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Benjamin Lerer (JIRA)" <ji...@apache.org> on 2016/02/09 14:01:18 UTC

[jira] [Commented] (CASSANDRA-10888) Tombstone error warning does not log partition key

    [ https://issues.apache.org/jira/browse/CASSANDRA-10888?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15138910#comment-15138910 ] 

Benjamin Lerer commented on CASSANDRA-10888:
--------------------------------------------

The patch for 2.1 looks fine but we need another one for 2.2 as the code has changed there.
It looks like what need to be done is to refactor {{TombstoneOverwhelmingException}} to pass the partitionKey to the constructor and remove the {{setKey}} method (It might worth double checking through).

If you could also check that the resulting patch apply to 3.0, it will be perfect.
Thanks

> Tombstone error warning does not log partition key
> --------------------------------------------------
>
>                 Key: CASSANDRA-10888
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-10888
>             Project: Cassandra
>          Issue Type: Improvement
>         Environment: DSE 4.8.2
>            Reporter: Ole Pedersen
>            Assignee: Brett Snyder
>         Attachments: 10888-2.1.txt
>
>
> Log partition key if read fails due to the ERROR threshold on read tombstoned cells.
> Right now I can specify a warning and an error threshold for C* when reading from a partition with many tombstones.
> If the query reads more than the “warning threshold” then C* writes a warning to the log with the partition key.
> But if a query reads more than the “error threshold” then C* aborts the query and writes to the log – but not the partition key, this time. 
> What I am missing is: Could C* also please write the partition key in case of query abort due to tombstones?



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