[
https://issues.apache.org/jira/browse/OOZIE-1519?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13758504#comment-13758504
]
Robert Kanter commented on OOZIE-1519:
--------------------------------------
To handle the HA propagation, the command can do what I did for the log
streaming where the REST command had an optional parameter, {{allservers}}.
When {{allserverse=false}} the receiving Oozie server will update its sharelib,
but not notify the other servers about it. When {{allservers=true}} or is
omitted (i.e. the default), the receiving Oozie server will update its
sharelib, and notify the other servers about it by calling the same command but
with {{alservers=false}} to prevent infinite propagation.
Separate logic shouldn't be necessary between HA and non-HA because the call to
get the other servers
({{Services.get().get(JobsConcurrencyService.class).getServerUrls()}}) will
return just the one server if HA is not enabled and all the servers if HA is
enabled.
> Admin command to update the sharelib
> ------------------------------------
>
> Key: OOZIE-1519
> URL: https://issues.apache.org/jira/browse/OOZIE-1519
> Project: Oozie
> Issue Type: Bug
> Reporter: Virag Kothari
>
> OOZIE-1518 aims to copy all the sharelibs from an hdfs staging dir. This JIRA
> will add a new admin command which will allow the admins to notify the oozie
> server when sharelib package is upgraded. In case of HA, one oozie server
> will notify other servers to update their own sharelibs. Also if the
> notification is lost or some server is down, that server will update its
> sharelib path during startup as it will check for any new sharelibs in the
> staging dir.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
