[
https://issues.apache.org/jira/browse/IGNITE-7606?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16353673#comment-16353673
]
Dmitriy Pavlov edited comment on IGNITE-7606 at 2/6/18 10:09 AM:
-----------------------------------------------------------------
https://github.com/apache/ignite/pull/3469 to be merged after IGNITE-7533
(https://github.com/apache/ignite/pull/3437)
was (Author: dpavlov):
https://github.com/apache/ignite/pull/3469 to be merged after
https://github.com/apache/ignite/pull/3437
> Write removed dirty page during replacement without holding segment write lock
> ------------------------------------------------------------------------------
>
> Key: IGNITE-7606
> URL: https://issues.apache.org/jira/browse/IGNITE-7606
> Project: Ignite
> Issue Type: Improvement
> Components: persistence
> Reporter: Dmitriy Pavlov
> Assignee: Dmitriy Pavlov
> Priority: Major
> Fix For: 2.5
>
> Attachments: chart_holdingLock.png, chart_notHoldingLock.png,
> chart_notHoldingLock2.png, putdumpAt17second.txt
>
>
> If a dirty page under the checkpoint is found, following is suggested
> - copy it to the local thread buffer,
> - and then after performing all actions in region for evicting the page
> - finish execution allocatePage()/acquirePage()
> - unlock segment to allow other workers to operate
> - perform the pwrite() call based on the data from local buffer
> Now if page eviction started there is possible drops to 0 put/seconds in case
> a lot of threads are watiting for same segment lock.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
