cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Ellis (JIRA)" <j...@apache.org>
Subject [jira] Commented: (CASSANDRA-1449) consistent nodetool blocking behavior
Date Mon, 20 Sep 2010 22:21:34 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-1449?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12912747#action_12912747
] 

Jonathan Ellis commented on CASSANDRA-1449:
-------------------------------------------

bq. Unless there's really any credit to having all the blocked calls go through the executor,
they should be invoked directly

absolutely there is a reason to go through the executor, many executors (commitlog, compaction,
probably others) avoid explicit locking because there is never supposed to be anyone but the
executor thread touching things.  apart from that, invoking from random jmx threads would
wreak havoc on our attempts to balance stage demand through executor pool sizing.

> consistent nodetool blocking behavior
> -------------------------------------
>
>                 Key: CASSANDRA-1449
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-1449
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Tools
>            Reporter: Brandon Williams
>            Assignee: Nirmal Ranganathan
>            Priority: Minor
>             Fix For: 0.7.0
>
>         Attachments: 0001-Updated-all-non-blocking-Nodetool-commands-to-blocki.patch
>
>
> Some operations in nodetool block.  Some don't.  We should choose a behavior and enforce
it, preferably blocking.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message