788147 - supervisor fails to reload itself on first installation

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Rail Aliiev [:rail]]

e.g.:
ue Sep 04 05:49:04 -0700 2012 /Stage[main]/Supervisord::Base/Service[supervisord] (err): Failed to call refresh: Could not restart Service[supervisord]: Execution of '/usr/bin/supervisorctl reload' returned 2:  at /etc/puppet/production/modules/supervisord/manifests/base.pp:47

That suggests either that the start is failing, or that puppet's not using "status" correctly.

Comment 1

[Dustin J. Mitchell [:dustin] (he/him)]

This is causing failures - albiet one-time failures - on all of the HP's right now.

Comment 2

[Dustin J. Mitchell [:dustin] (he/him)]

> info: /Stage[main]/Supervisord::Base/File[/etc/supervisord.conf]: Scheduling refresh of Service[supervisord]
> debug: Service[supervisord](provider=redhat): Executing '/sbin/service supervisord status'
> debug: Puppet::Type::Service::ProviderRedhat: Executing '/sbin/chkconfig supervisord'
> debug: Service[supervisord](provider=redhat): Executing '/sbin/service supervisord start'
> debug: Puppet::Type::Service::ProviderRedhat: Executing '/sbin/chkconfig supervisord'
> debug: Puppet::Type::Service::ProviderRedhat: Executing '/sbin/chkconfig supervisord on'
> notice: /Stage[main]/Supervisord::Base/Service[supervisord]/ensure: ensure changed 'stopped' to 'running'
> debug: /Stage[main]/Supervisord::Base/Service[supervisord]: The container Class[Supervisord::Base] will propagate my refresh event
> debug: Service[supervisord](provider=redhat): Executing '/sbin/service supervisord status'
> debug: Service[supervisord](provider=redhat): Executing '/usr/bin/supervisorctl reload'
> err: /Stage[main]/Supervisord::Base/Service[supervisord]: Failed to call refresh: Could not restart Service[supervisord]: Execution of '/usr/bin/supervisorctl reload' returned 2:  at /etc/puppet/production/modules/supervisord/manifests/base.pp:47

So it is correctly starting it, but the reload isn't working:

[root@relabs08 ~]# /usr/bin/supervisorctl reload
error: <class 'socket.error'>, [Errno 2] No such file or directory: file: <string> line: 1

Possibly related:

[root@relabs08 ~]# supervisorctl 
unix:///var/tmp/supervisor.sock no such file

So I think this is not preventing production right now, but will hurt next time you change supervisord settings, as it won't reload correctly.

I suspect the problem is in supervisord.conf -- the socket should be specified in [unix_http_server]:file, not [supervisord]:http_port

Comment 3

[Chris AtLee [:catlee]]

that parameter changed names between supervisord 2.x and 3.x

Comment 4

[Rail Aliiev [:rail]]

Attachment: supervisord config

I managed to connect to the supervisord server via supervisorctl with this change.

Comment 5

[Rail Aliiev [:rail]]

Comment on attachment 658152 [details] [diff] [review]
supervisord config

http://hg.mozilla.org/build/puppet/rev/7e8fd836f973