[
https://issues.apache.org/jira/browse/GEODE-9753?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17439191#comment-17439191
]
ASF GitHub Bot commented on GEODE-9753:
---------------------------------------
gaussianrecurrence opened a new pull request #891:
URL: https://github.com/apache/geode-native/pull/891
- While serializing a PdxSerializable object, there is a possible race
condition which might cause the client to crash.
This race-condition happens whenever the cluster is restarted during
the serialization process and if
on-client-disconnect-clear-pdxType-Ids is set to true, meaning the
PdxTypeRegistry will be cleaned up if the connection towards the
cluster is lost.
- This issue has been solved by using the previously fetched local PDX
type.
- In order to properly test this solution, PdxRemoteWriterFactory has been
added,
so the race-condition can be force at test-time.
- A new IT has been added to test that the solution is working fine.
- make_unique was needed inside cppcache/src, so it was moved to
utils/cxx_extensions.hpp
- Also in order to ease the use of newer C++ standard a preprocessor
check was added to make_unique, so if the used standard >= C++14, the
standard implementation of make_unique will be used instead.
--
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.
To unsubscribe, e-mail: notifications-unsubscr...@geode.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> Coredump during PdxSerializable object serialization
> ----------------------------------------------------
>
> Key: GEODE-9753
> URL: https://issues.apache.org/jira/browse/GEODE-9753
> Project: Geode
> Issue Type: Bug
> Components: native client
> Reporter: Mario Salazar de Torres
> Assignee: Mario Salazar de Torres
> Priority: Major
> Labels: needsTriage
>
> *GIVEN* **a single server and locator cluster with 1 PdxSerializable class
> registered, named Order
> *AND* a geode-native client with 1 PdxSerializable 1 PdxSerializable class
> registered, named Order
> *WHEN* a Order object is tried to be serialized
> *WHILE* the cluster is being restarted
> *THEN* a coredump happens due to PdxType=nullptr
> —
> *Additional information*. As seen by early troubleshooting, the coredump
> happens because the pdx type is tried to be fetched from the PdxTypeRegist by
> its class name, but the PdxTypeRegistry is cleaned up during serialization
> given that subscription redundancy was lost.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
