[
https://issues.apache.org/jira/browse/HDFS-12443?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16216760#comment-16216760
]
Weiwei Yang commented on HDFS-12443:
------------------------------------
Thanks for taking over [~linyiqun]. Some more background, currently the code
uses _push_ mode, which means SCM collects TXs and caches them for DNs. Instead
of _pull_ (DN fetches TXs from SCM while doing HB processing). This was because
we want to make HB lightweight as much as possible (I/O-less). But this
approach makes it a bit difficult to prepare a reasonable size of TXs for DNs,
please take a look at this issue and let me know if you have any idea to
resolve it. Thanks!
> Ozone: Improve SCM block deletion throttling algorithm
> -------------------------------------------------------
>
> Key: HDFS-12443
> URL: https://issues.apache.org/jira/browse/HDFS-12443
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: ozone, scm
> Reporter: Weiwei Yang
> Assignee: Yiqun Lin
> Labels: OzonePostMerge
>
> Currently SCM scans delLog to send deletion transactions to datanode
> periodically, the throttling algorithm is simple, it scans at most
> {{BLOCK_DELETE_TX_PER_REQUEST_LIMIT}} (by default 50) at a time. This is
> non-optimal, worst case it might cache 50 TXs for 50 different DNs so each DN
> will only get 1 TX to proceed in an interval, this will make the deletion
> slow. An improvement to this is to make this throttling by datanode, e.g 50
> TXs per datanode per interval.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
