[ 
https://issues.apache.org/jira/browse/CASSANDRA-18627?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17737322#comment-17737322
 ] 

Josh McKenzie edited comment on CASSANDRA-18627 at 6/26/23 9:24 PM:
--------------------------------------------------------------------

-Hm. To anyone wandering past and curious, I'm not convinced this actually 
_fixes_ anything since the only call sites in the {{PendingRepairManager}} 
actually don't pass in any additional sstables or bytes (just 0), so the 
compaction strategies don't have any estimated extra sstables or bytes to 
calculate anything from. I'll need to dig a bit into where those values make 
their way into this scope to then be passed on to compaction to determine where 
and why they're vestigial in this way.-

Restructured ticket to clean this code up since it's clearly confusing. :)


was (Author: jmckenzie):
Hm. To anyone wandering past and curious, I'm not convinced this actually 
_fixes_ anything since the only call sites in the {{PendingRepairManager}} 
actually don't pass in any additional sstables or bytes (just 0), so the 
compaction strategies don't have any estimated extra sstables or bytes to 
calculate anything from. I'll need to dig a bit into where those values make 
their way into this scope to then be passed on to compaction to determine where 
and why they're vestigial in this way.

> Tidy up vestigial code left over after CASSANDRA-17931
> ------------------------------------------------------
>
>                 Key: CASSANDRA-18627
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-18627
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Consistency/Repair, Local/Compaction
>            Reporter: Josh McKenzie
>            Assignee: Josh McKenzie
>            Priority: Normal
>             Fix For: 5.x
>
>
> In CASSANDRA-17931, some code was left over (\{{additionalBytes}}, 
> {{{}additionalSSTables{}}}) inside the {{PendingRepairManager}} which, while 
> not explicitly hurting anything, provides for some confusion around when we 
> do or don't take additional values into account when adjusting for pending 
> operations when deciding what to do with upcoming repairs etc.
>  
> We should tidy that up and consider making the role of those methods / their 
> structure more clear between the \{{PendingRepairManager]} hierarchy and the 
> {{AbstractCompactionStrategy}} hierarchy to prevent further confusion.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

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

Reply via email to