[
https://issues.apache.org/jira/browse/IGNITE-26132?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Kirill Tkalenko updated IGNITE-26132:
-------------------------------------
Description:
I noticed that in some unfortunate circumstances we can see a lot of
*Checkpoint pages were not written yet due to unsuccessful page write lock
acquisition and will be retried [pageCount=1]* messages in the log, I think we
should reduce the number of such messages.
We can fix this, for example, when a certain number of attempts to retry write
dirty pages is reached, take the lock on the page not by attempting but with
blocking waiting.
was:
I noticed that in some unfortunate circumstances we can see a lot of
"Checkpoint pages were not written yet due to unsuccessful page write lock
acquisition and will be retried [pageCount=1]" messages in the log, I think we
should reduce the number of such messages.
We can fix this, for example, when a certain number of attempts to retry write
dirty pages is reached, take the lock on the page not by attempting but with
blocking waiting.
> Improve retry writing of dirty pages
> ------------------------------------
>
> Key: IGNITE-26132
> URL: https://issues.apache.org/jira/browse/IGNITE-26132
> Project: Ignite
> Issue Type: Improvement
> Reporter: Kirill Tkalenko
> Assignee: Kirill Tkalenko
> Priority: Major
> Labels: ignite-3
>
> I noticed that in some unfortunate circumstances we can see a lot of
> *Checkpoint pages were not written yet due to unsuccessful page write lock
> acquisition and will be retried [pageCount=1]* messages in the log, I think
> we should reduce the number of such messages.
> We can fix this, for example, when a certain number of attempts to retry
> write dirty pages is reached, take the lock on the page not by attempting but
> with blocking waiting.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
