[
https://issues.apache.org/jira/browse/GEODE-8793?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17291003#comment-17291003
]
ASF GitHub Bot commented on GEODE-8793:
---------------------------------------
gaussianrecurrence edited a comment on pull request #715:
URL: https://github.com/apache/geode-native/pull/715#issuecomment-785990080
Since it has passed some time since I last updated this PR here is a little
update:
- I've changed the behaviour of PdxTypeRegistry cleanup. Now is cleaned up
whenever the are no more alive endpoints in the pool, rather than whenever
connections are exhausted. This was changed because if the cluster is restarted
and no requests were being executed meanwhile it could happen that the registry
was not cleaned up. But, on the other hand endpoint liveness is controlled by
periodic ping requests which are regularly sent.
- I've noticed a edge case in which the client can crash if the registry is
cleaned up while PdxHelper::serializePdx is being called. I've created an issue
to tackle that problem: https://issues.apache.org/jira/browse/GEODE-8968
- I've noticed a quite concerning bug having to do with
throwExceptionIfError, which I solved for the specific case but I created a new
Jira ticket to tackle all the cases:
https://issues.apache.org/jira/browse/GEODE-8973
----------------------------------------------------------------
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
> Fix PdxTypeRegistry cleanup upon cluster disconnection
> ------------------------------------------------------
>
> Key: GEODE-8793
> URL: https://issues.apache.org/jira/browse/GEODE-8793
> Project: Geode
> Issue Type: Bug
> Components: native client
> Affects Versions: 1.13.0, 1.13.1
> Reporter: Mario Salazar de Torres
> Assignee: Mario Salazar de Torres
> Priority: Major
> Labels: pull-request-available
>
> *GIVEN* A native client cache with on-client-disconnect-clear-pdxType-Ids is
> set to "true" (not the best name if you ask me)
> *WHEN* the connection towards the cluster is lost
> *THEN* PdxTypeRegistry should be cleaned up straight away.
> ----
> *Additional information*
> **Instead of being cleaned up straight away, PdxTypeRegistry it's cleaned up
> every 2 cluster disconnections.
> Probably the issue is that semaphores in place are missused.
> Consider additionally to change the callback name "cliCallback" as it does
> not represent the actual functionality here.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
