[jira] [Commented] (GEODE-4730) Remove DLock acquisition timeouts in ClusterConfigurationService.createConfigurationResponse

["ASF subversion and git services (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/GEODE-4730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16386191#comment-16386191 ] 

ASF subversion and git services commented on GEODE-4730:
--------------------------------------------------------

Commit e5025d399c56b1f86e36c37638108d27655886af in geode's branch refs/heads/develop from [~jens.deppe]
[ https://gitbox.apache.org/repos/asf?p=geode.git;h=e5025d3 ]

GEODE-4765: Add retries when retrieving cluster config (#1539)

- This helps when a locator is reconnecting to the distributed system and isn't
  fully ready yet.
- During reconnect, pass a fresh cache to the ClusterConfigurationService -
  this fixes the 'DLock destroyed' issue.
- Also fixes GEODE-4730


> Remove DLock acquisition timeouts in ClusterConfigurationService.createConfigurationResponse
> --------------------------------------------------------------------------------------------
>
>                 Key: GEODE-4730
>                 URL: https://issues.apache.org/jira/browse/GEODE-4730
>             Project: Geode
>          Issue Type: Improvement
>          Components: configuration
>            Reporter: Jens Deppe
>            Assignee: Jens Deppe
>            Priority: Major
>
> Reviewing some of this code with [~upthewaterspout] it seems like it would be worse if the lock acquisition fails and an empty response is returned vs. a hang/deadlock happening on server startup.
> At least if there's a hang we'd know it as opposed to a server starting up without any configuration.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)