[
https://issues.apache.org/jira/browse/OOZIE-3183?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16364068#comment-16364068
]
Peter Cseh commented on OOZIE-3183:
-----------------------------------
+1
Improving logging is always a good idea. Also Oracle's documentation of
waitFor() states that it uses exitCode() by default, but it should be improved
by the native implementations.
> Better logging for SshActionExecutor and extended HA capability when calling
> to remote host
> -------------------------------------------------------------------------------------------
>
> Key: OOZIE-3183
> URL: https://issues.apache.org/jira/browse/OOZIE-3183
> Project: Oozie
> Issue Type: Improvement
> Components: action
> Affects Versions: 5.0.0b1
> Reporter: Andras Piros
> Assignee: Andras Piros
> Priority: Major
> Fix For: 5.0.0
>
> Attachments: OOZIE-3183.001.patch
>
>
> In some High Availability environments, when an SSH action's
> {{SshActionExecutor#start()}} runs on another node than
> {{SshActionExecutor#check()}}, it can happen that due to network bandwidth
> reasons calling {{Process#exitValue()}} is not enough - but calling
> {{Process#waitFor()}} is better.
> Also, for better error root cause analysis, enhance logging along
> {{SshActionExecutor}} for the critical parts.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
