cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Parth Setya (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (CASSANDRA-9776) Tombstone Compaction Not Getting Triggered on a single SSTable even when the % of Droppable > 85%
Date Tue, 14 Jul 2015 05:39:05 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-9776?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14625782#comment-14625782
] 

Parth Setya edited comment on CASSANDRA-9776 at 7/14/15 5:38 AM:
-----------------------------------------------------------------

[Edit: I also added unchecked_tombstone_compaction as true. No impact.]

I started a stable write load after major compaction. However, I do have some considerations
here:
1. I set the value of tombstone_compaction_interval to 0 (assuming this does not disable Tombstone
Compaction)
2. Also, I started the writes *after* the single large sstables (containing > 85% tombstones)
was created. 




was (Author: tehkidnextdoor):
[Edit: I also added *unchecked_tombstone_compaction* as true. No impact.]

I started a stable write load after major compaction. However, I do have some considerations
here:
1. I set the value of tombstone_compaction_interval to 0 (assuming this does not disable Tombstone
Compaction)
2. Also, I started the writes *after* the single large sstables (containing > 85% tombstones)
was created. 



> Tombstone Compaction Not Getting Triggered on a single SSTable even when the % of Droppable
> 85%
> -------------------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-9776
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9776
>             Project: Cassandra
>          Issue Type: New Feature
>         Environment: RHEL
> Apache Cassandra v 2.0.14
>            Reporter: Parth Setya
>             Fix For: 2.0.14
>
>
> The following steps can be used to Replicate the Issue:
> 1. Reduce gc_grace_seconds to 18000
> 2. Set tombstone_compaction_interval to 1 day
> 3. Insert 9 Million Rows
> 4. Delete 9 Million Rows
> 5. Insert 1 Million Rows
> 6. Run Major Compaction so that the Tombstones for 9 million rows and expiring columns
for 1 million rows are in the same SStable.(In may case the size of the resultant sstable
was 963M)
> Note: I also ran "nodetool cfstats" command and the found that the number of keys in
the CF to be 10000000(as expected)
> Tombstone Compaction on the resultant SSTable should have been triggered after 1 day
after the creation of that Table but nothing happened for > 2 days.
> After 3 days I ran major compaction the size of the table was reduced to 213M.



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

Mime
View raw message