Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Kafka >> mail # dev >> [GitHub] kafka pull request: KAFKA-1414: Speedup broker startup after hard ...


Copy link to this message
-
[GitHub] kafka pull request: KAFKA-1414: Speedup broker startup after hard ...
GitHub user ataraxer opened a pull request:

    https://github.com/apache/kafka/pull/26

    KAFKA-1414: Speedup broker startup after hard reset and shutdown

    This patch increases speed of both hard reset and shutdown by introducing `log.recovery.threads` and `log.shutdown.threads` properties, which allows to perform work required for them in parallel, grained by log directories.
    
    Best performance can be achieved by setting thread count to number of log directories, provided that they are located on dedicated drives. Although that option should be used with caution due to the possibility of native JVM out of memory error.
    
    Patch is compiled of changes proposed by Jay Kreps, Alexey Ozeritskiy, Dmitry Bugaychenko by Anton Karamanov.
    
    All tests are passing.

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/ataraxer/kafka kafka-1414

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/kafka/pull/26.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #26
    
commit e4a86709d07030c44f077ab20d4329ddb84c4aec
Author: Anton Karamanov <[EMAIL PROTECTED]>
Date:   2014-07-15T17:42:15Z

    KAFKA-1414 Speedup broker startup after hard reset and shutdown; patched by Jay Kreps,  Alexey Ozeritskiy, Dmitry Bugaychenko and Anton Karamanov

If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [EMAIL PROTECTED] or file a JIRA ticket
with INFRA.