[
https://issues.apache.org/jira/browse/IGNITE-1377?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14933335#comment-14933335
]
Semen Boikov commented on IGNITE-1377:
--------------------------------------
Now metadata update will hang if it is updated inside user transaction (or
atomic update) and metadata primary node fails. Need to change metadata update
code to do not use automatic cache updated retries, but throw some exception
instead. Need review all places where IgniteCacheObjectProcessor's methods
toCacheKeyObject/toCacheObject are used and add retry logic. Also need add more
tests for metadata update retry logic.
> Portable metadata update on changing topology causes hangs in tests
> -------------------------------------------------------------------
>
> Key: IGNITE-1377
> URL: https://issues.apache.org/jira/browse/IGNITE-1377
> Project: Ignite
> Issue Type: Sub-task
> Reporter: Alexey Goncharuk
> Assignee: Semen Boikov
> Fix For: ignite-1.5
>
>
> I removed metadata update on for objects not defined in configuration (this
> is not an introduced bug, it existed before) because it causes portable
> failover suite to hang.
> Need to investigate further the cause of the hang and unmute
> {{testNoConfiguration}}.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
