[
https://issues.apache.org/jira/browse/GEODE-2863?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16846987#comment-16846987
]
Mario Ivanac edited comment on GEODE-2863 at 5/23/19 7:17 PM:
--------------------------------------------------------------
My proposal is, that all task that are in pooll, at the moment close/destroy
request is received, will be finished, and all new tasks received from that
moment, to be discarded.
was (Author: mivanac):
My proposal is, that all task that are in pull, at the moment close/destroy
request is received, will be finished, and all new tasks received from that
moment, to be discard.
> DiskStoreImpl close does not shutdown its thread pools
> ------------------------------------------------------
>
> Key: GEODE-2863
> URL: https://issues.apache.org/jira/browse/GEODE-2863
> Project: Geode
> Issue Type: Bug
> Components: persistence
> Affects Versions: 1.0.0-incubating
> Reporter: Darrel Schneider
> Assignee: Mario Ivanac
> Priority: Major
> Labels: storage_3
>
> Closing a DiskStoreImpl does not shutdown its thread pools.
> It has two pools, "diskStoreTaskPool" and "delayedWritePool", that it does
> not shutdown.
> However at least in some cases (see GEODE-2862) it does wait for all the
> tasks submitted to these pools to complete. But it is not clear if additional
> tasks could still be submitted after this wait is done.
> Some logic should be added that causes the code that would submit a task to
> these executors to no longer do so or, if needed, to process the task inline.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
