[
https://issues.apache.org/jira/browse/GEODE-8513?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17204131#comment-17204131
]
ASF GitHub Bot commented on GEODE-8513:
---------------------------------------
pivotal-jbarrett commented on pull request #5532:
URL: https://github.com/apache/geode/pull/5532#issuecomment-700840958
Two interesting things were happing in there. First was the serialization of
the session to disk, which really only made sense in peer mode since in client
server the persistence is off JVM already. In peer mode it makes more sense,
and is more tested, to use region persistence to achieve this hot deployment
behavior. The second thing going on was the removal from the local region. This
was to avoid class loader casting issues. This was just simplified to a local
clear operation rather than iterating over all the sessions specific to this
web context. The upside is that it takes a lot less time and doesn't invoke any
queries. The downside is that other web contexts will have to refresh their
local caches too but this scenario should be very small.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> DeltaSesssionManager writes all sessions to disk on shutdown
> ------------------------------------------------------------
>
> Key: GEODE-8513
> URL: https://issues.apache.org/jira/browse/GEODE-8513
> Project: Geode
> Issue Type: Bug
> Components: http session
> Reporter: Jacob Barrett
> Assignee: Jacob Barrett
> Priority: Major
> Labels: pull-request-available
>
> DeltaSesssionManager executes a query to fetch all sessions for the current
> context shutting down and then serializes them to disk. There is no need to
> do this since the sessions are persisted in the session region on the
> servers.
> This leads:
> * to unaccounted for disk usage.
> * execution of a potentially large query at shutdown.
> * class casting executions in antlr in some Tomcat instances.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
