[
https://issues.apache.org/jira/browse/IGNITE-5795?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16651182#comment-16651182
]
Ignite TC Bot commented on IGNITE-5795:
---------------------------------------
{panel:title=Possible
Blockers|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1}
{color:#d04437}Hadoop{color} [[tests 0 TIMEOUT , Out Of Memory Error , Exit
Code , [Artifacts publishing failed]
|https://ci.ignite.apache.org/viewLog.html?buildId=2094928]]
{color:#d04437}_Javadoc_{color} [[tests 0 Exit Code , BuildFailureOnMessage
|https://ci.ignite.apache.org/viewLog.html?buildId=2094952]]
{panel}
[TeamCity Run
All|http://ci.ignite.apache.org/viewLog.html?buildId=2087699&buildTypeId=IgniteTests24Java8_RunAll]
> Binary metadata is not registered during start of cache
> -------------------------------------------------------
>
> Key: IGNITE-5795
> URL: https://issues.apache.org/jira/browse/IGNITE-5795
> Project: Ignite
> Issue Type: Bug
> Components: sql
> Affects Versions: 2.0
> Reporter: Dmitry Karachentsev
> Assignee: Anton Kalashnikov
> Priority: Major
> Labels: usability
> Fix For: 2.8
>
>
> When cache configured with QueryEntity and used key type with
> @AffinityKeyMapped field, it will be ignored and wrong partition calculated.
> This happens because QueryEntity processing precedes key type registering in
> binary meta cache. On that step
> CacheObjectBinaryProcessorImpl#affinityKeyField called and unable to resolve
> type, so null returned and null putted in affKeyFields.
> On next put/get operation CacheObjectBinaryProcessorImpl#affinityKeyField
> will return null from affKeyFields, but should be affinity key field.
> Test that reproduces problem in [PR
> 2330|https://github.com/apache/ignite/pull/2330]
> To wrorkaround the issue, set IgniteConfiguration#setKeyConfiguration(), it
> will force registering key.
> *Root cause of this behaviour* is that in during cache registration
> registration of metadata doesn't happen. It lead to exchange messages like
> (MappingProposedMessage, MetadataRequestMessage etc.) which are often
> redundant and also lead to behaviour described above(unhandled
> AffinityKeyMapped).
> Main idea to fix this problem it is registration of metadata locally in
> during cache start on each node without messages exchange.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
