You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Corentin Chary (JIRA)" <ji...@apache.org> on 2016/12/13 22:43:58 UTC

[jira] [Comment Edited] (CASSANDRA-13040) Estimated TS drop-time histogram updated with Cell.NO_DELETION_TIME

    [ https://issues.apache.org/jira/browse/CASSANDRA-13040?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15746508#comment-15746508 ] 

Corentin Chary edited comment on CASSANDRA-13040 at 12/13/16 10:43 PM:
-----------------------------------------------------------------------

Additional fix related to this in CASSANDRA-13038, looks like this patch still update the tombstone histogram for cells with a ttl and this has a significant performance penalty.


was (Author: iksaif):
Additional fix related to this in CASSANDRA-13038, looks like this patch still update the histogram for cells with a ttl.

> Estimated TS drop-time histogram updated with Cell.NO_DELETION_TIME
> -------------------------------------------------------------------
>
>                 Key: CASSANDRA-13040
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-13040
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Stefan Podkowinski
>            Assignee: Stefan Podkowinski
>             Fix For: 3.0.11, 3.10, 3.12, 4.0
>
>         Attachments: 13040-3.0.patch
>
>
> Starting with 3.0, {{MetadataCollector.updateLocalDeletionTime(int newLocalDeletionTime)}} will be called with {{cell.localDeletionTime()}} in order to update the min/max deletion times as well as the {{estimatedTombstoneDropTime}} histogram. Currently this also happens for {{Cell.NO_DELETION_TIME}} ({{Integer.MAX_VALUE}}), which causes the histogram to be called for every regular cell.
> This can be easily verified by using {{sstablemetadata}} on any sstable. You should be able to find a very high value for 2147483647 in the "Estimated tombstone drop times" section.
> At this point, I don't think this is causing serious harm, but could be a performance issue (see linked ticket). Calculated droppable TS times should not be affected.



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