[jira] [Commented] (CURATOR-508) LeaderLatch#setNode might throws unexpected NoNodeException

["TisonKun (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CURATOR-508?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16774249#comment-16774249 ] 

TisonKun commented on CURATOR-508:
----------------------------------

Isn't such case *inside Curator*?

Specifically, Curator suffers from unstable ZK connection and ZK deletes the EPHEMERAL node(outPath). Later, Curator reconnected to ZK and try to {{#reset}}, and then tried to delete the previous node which had been deleted by ZK.

Or we never meet this scenario?

> LeaderLatch#setNode might throws unexpected NoNodeException
> -----------------------------------------------------------
>
>                 Key: CURATOR-508
>                 URL: https://issues.apache.org/jira/browse/CURATOR-508
>             Project: Apache Curator
>          Issue Type: Bug
>          Components: Recipes
>            Reporter: TisonKun
>            Priority: Major
>
> {code:java}
> private void setNode(String newValue) throws Exception
> {
>  String oldPath = ourPath.getAndSet(newValue);
>  if ( oldPath != null )
>  {
>  client.delete().guaranteed().inBackground().forPath(oldPath);
>  }
> }{code}
> What if the election node deleted for some reason, e.g., the session disconnected and reconnected? Should we use {{#quietly()}} here?



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