You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@bigtop.apache.org by "Chunjun Xiao (JIRA)" <ji...@apache.org> on 2014/11/03 10:15:34 UTC
[jira] [Comment Edited] (BIGTOP-1500) Stop already stopped hue
service returns ok instead of failed
[ https://issues.apache.org/jira/browse/BIGTOP-1500?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14194187#comment-14194187 ]
Chunjun Xiao edited comment on BIGTOP-1500 at 11/3/14 9:15 AM:
---------------------------------------------------------------
I'm fine with providing the patch.
The problem is, how would we want to unify the behavior of service stopping?
When stop a stopped service, should we log_success_msg or log_failure_msg?
For most services (e.g., hadoop, hbase, hive, etc.), log_success_msg is called.
But for some service (e.g., flume-agent), log_failure_msg is called, same as nfsd/httpd.
All the above services returns 0.
So, do we really want to change the behavior of most services to comply with linux service convention?
was (Author: chunjun.xiao):
Sure. I will need to check other components to make sure the service stop behavior of all components is consistent.
> Stop already stopped hue service returns ok instead of failed
> -------------------------------------------------------------
>
> Key: BIGTOP-1500
> URL: https://issues.apache.org/jira/browse/BIGTOP-1500
> Project: Bigtop
> Issue Type: Improvement
> Components: build
> Affects Versions: 0.8.0
> Reporter: Chunjun Xiao
> Priority: Trivial
>
> Generally, when we stop some service already stopped, we will get an error return, as nfsd/httpd does.
> {quote}
> [root@svm7 /etc/init.d] # service nfs stop
> Shutting down NFS daemon: [ OK ]
> Shutting down NFS mountd: [ OK ]
> Shutting down NFS quotas: [ OK ]
> [root@svm7 /etc/init.d] # service nfs stop
> Shutting down NFS daemon: [FAILED]
> Shutting down NFS mountd: [FAILED]
> Shutting down NFS quotas: [FAILED]
> [root@svm7 /etc/init.d] #
> {quote}
> While when we stop hue (and many other hadoop components), which has already been stopped, it will return success.
> {quote}
> [root@svm7 /etc/init.d] # service hue stop
> [root@svm7 /etc/init.d] # [ OK ]
> [root@svm7 /etc/init.d] # service hue stop
> [root@svm7 /etc/init.d] # [ OK ]
> {quote}
> Should the hadoop services comply with the convention?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)