[
https://issues.apache.org/jira/browse/IGNITE-4022?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16466009#comment-16466009
]
Denis Mekhanikov commented on IGNITE-4022:
------------------------------------------
With this approach we will get such warnings for every reassignment and every
service, that wasn't assigned to any nodes. It will flood the log.
I think, we should print such warning in two cases:
* This is the first time the assignments are calculated, i.e. {{oldAssigns ==
null}}
* Old assignments are not empty, i.e. there was at least one assigned node
after last reassignment.
In this case warnings will be printed only during first assignment and when
nodes with last working service instances are brought down.
> IgniteServices soesn't throw an exception if there are no server nodes
> ----------------------------------------------------------------------
>
> Key: IGNITE-4022
> URL: https://issues.apache.org/jira/browse/IGNITE-4022
> Project: Ignite
> Issue Type: Bug
> Affects Versions: 1.8
> Reporter: Alexander Belyak
> Assignee: Vladislav Pyatkov
> Priority: Major
>
> If you call deployNodeSingleton method, but there are no server nodes in
> IgniteServices base ClusterGroup - you will never know about it and can't
> find deployed service instance. Probably we should print out these errors in
> logs as well.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
