[
https://issues.apache.org/jira/browse/GEODE-9346?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17378211#comment-17378211
]
ASF subversion and git services commented on GEODE-9346:
--------------------------------------------------------
Commit 11e4c3a4ca4bf7ef2203e0fdd111e536e5721e50 in geode's branch
refs/heads/develop from Xiaojian Zhou
[ https://gitbox.apache.org/repos/asf?p=geode.git;h=11e4c3a ]
GEODE-9346: When client received incorrect byte array of PdxType due … (#6561)
> When client received incorrect byte array of PdxType due to broken socket, it
> should be retried
> ------------------------------------------------------------------------------------------------
>
> Key: GEODE-9346
> URL: https://issues.apache.org/jira/browse/GEODE-9346
> Project: Geode
> Issue Type: Bug
> Affects Versions: 1.12.2, 1.13.2, 1.14.0, 1.15.0
> Reporter: Xiaojian Zhou
> Assignee: Xiaojian Zhou
> Priority: Major
> Labels: pull-request-available
>
> Client's query for PdxType will get a byte array in response message. The
> byte array is the same at the server, but different query from different
> client might receive wrong byte array and end up with
> PdxSerializationException in scalability test with server HA. This could
> caused by socket broken and bytes are not flushed. We expected such broken
> socket scenario and prepared some retry lock. However, our retry logic did
> not consider above scenario, i.e. the message header is correct, but the
> embedded byte array for PdxType is wrong.
> The solution is to retry in this case.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
