[
https://issues.apache.org/jira/browse/IGNITE-3808?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15529766#comment-15529766
]
Andrey Gura commented on IGNITE-3808:
-------------------------------------
Can' reproduce issue with this reproducer. I tried on versions 1.5, 1.6 and 1.7.
The problem is very weird for me because even if {{readFromBackup == true}} and
synchronization is {{PRIMARY_SYNC}} each invoke of entry processor will be
performed on primary node. The only situation when described problem is
possible if both primary and backup nodes are gone and partition is lost. In
this case {{cache.get()}} also should return {{null}}. Also you can listen
{{EVT_CACHE_REBALANCE_PART_DATA_LOST}} event on all nodes (should be
recordable) that should be triggered in case of all owner nodes are gone.
> Missing EntryProcessor invocations with partitions in flight
> ------------------------------------------------------------
>
> Key: IGNITE-3808
> URL: https://issues.apache.org/jira/browse/IGNITE-3808
> Project: Ignite
> Issue Type: Bug
> Components: cache
> Affects Versions: 1.7
> Reporter: Patrick Peralta
> Priority: Critical
>
> With {{CacheAtomicityMode.ATOMIC}}, on occasion we see entry processor
> invocations that go missing when partitions are in flight.
> Reproducer will follow in a PR.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
