Puppet users - Nov 2012 - Apparent problem with resource ordering using Puppetlabs Corosync module (does not bring up HA resources in correct order when configured to do so)

I''m setting up a simple application management with corosync via the 
puppetlabs module. 

What SHOULD happen is: 

1. Ip gets brought up
2. jboss gets brought up

I have defined the resources like this: 

    cs_primitive { ''applayer_ip'':
      primitive_class     => ''ocf'',
      primitive_type      => ''IPaddr2'',
      provided_by         => ''heartbeat'',
      parameters          => {
        ''ip'' => "$servicelayer::jbossip",
        ''cidr_netmask'' =>
"$servicelayer::jbossmask",
      },
      operations          => { ''monitor'' => {
''interval'' => ''10s'' } },
      ensure              => ''present'',
    }

    cs_primitive { ''applayer_jboss'':
      primitive_class     => ''lsb'',
      primitive_type      => ''jboss'',
      provided_by         => ''heartbeat'',
      operations          => {
        ''monitor''         => { ''interval''
=> ''1min'', ''timeout'' =>
''30s'' },
        ''start''           => { ''interval''
=> ''30s'', ''timeout'' =>
''30s'',
''on-fail'' => ''restart'' },
        ''stop''            => { ''interval''
=> ''30s'', ''timeout'' =>
''30s'',
''on-fail'' => ''restart'' },
      },
      require             => Cs_primitive[''applayer_ip''],
      metadata            => {''target-role'' =>
"Started"},
      ensure              => present,
    }
    cs_colocation { ''jboss_with_ip'':
      primitives          => [ ''applayer_ip'',
''applayer_jboss'' ],
    }

    cs_order { ''jboss_ip_before_jboss'':
    first   => ''applayer_ip'',
    second  => ''applayer_jboss'',
    require => Cs_colocation[''jboss_with_ip''],
    }

The puppet run succeeds and appears to apply everything without any errors. 

However, when I check the status of the cluster, jboss is up and running 
while the IP is not: 

===========Last updated: Fri Nov 30 21:46:51 2012
Last change: Fri Nov 30 21:28:41 2012 via cibadmin on test1.prod1
Stack: openais
Current DC: test1.prod1 - partition WITHOUT quorum
Version: 1.1.6-3.el6-a02c0f19a00c1eb2527ad38f146ebc0834814558
1 Nodes configured, 2 expected votes
1 Resources configured.
===========
Online: [ test1.prod1 ]

applayer_jboss  (lsb:jboss):    Started test1.prod1 (unmanaged) FAILED

Failed actions:
    applayer_jboss_stop_0 (node=test1.prod1, call=4, rc=1, 
status=complete): unknown error

Unfortunately I am not fantastically experienced with either puppet or 
corosync, so it''s entirely possible that I''m doing something
silly - but if
I manually add the IP on the command line, corosync brings it up and 
manages it normally (with the same arguments exactly). So it seems like the 
module is doing something funny here. 

Help?





-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/puppet-users/-/Nec9BWaHmkwJ.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

Which puppetlabs-corosync are you using?  There are 19 forks of it last I 
checked(including one by me for my own needs surrounding DRBD which is now 
horribly out of date).

I would change this:
 cs_order { ''jboss_ip_before_jboss'':
    first   => ''applayer_ip'',
    second  => ''applayer_jboss'',
    require => Cs_colocation[''jboss_with_ip''],
    }
to
 cs_order { ''jboss_ip_before_jboss'':
    first   => ''applayer_ip'',
    second  => ''applayer_jboss'',
    score => ''INFINITY'',
    require =>
Cs_primitive[''applayer_ip'',''applayer_jboss''],
    }
and 
cs_colocation { ''jboss_with_ip'':
      primitives          => [ ''applayer_ip'',
''applayer_jboss'' ],
      require =>
Cs_primitive[''applayer_ip'',''applayer_jboss''],
    }

as autorequires were very, very dicey when I mucked with the module. 
 Ignore this if they work now.

Also - make sure you''re not setting up the IP or the service anywhere
else
as corosync should be managing it.

Outside of puppet, are you sure IPaddr2 is the right primitive_class?  Can 
we get the output of a crm configure show, and if it''s missing the 
primitive if it takes:
crm configure primitive applayer_ip ocf:heartbeat:IPaddr2 params 
ip="whatever" cidr_netmask="whatever"

because I suspect that''s the problem.

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/puppet-users/-/tzNfECQp_LcJ.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

Hi, Nick - 

Thanks for the response!

Those changes do appear to have made the manifest apply and work correctly. 
 Now I just have to figure out why corosync can''t get jboss to stop
once it
has been started (I can''t either, ''service jboss
stop'' seems to hang
indefinitely). 

On to the next challenge...



On Saturday, December 1, 2012 11:20:06 AM UTC-7, Nick Cammorato
wrote:
>
> Which puppetlabs-corosync are you using?  There are 19 forks of it last I 
> checked(including one by me for my own needs surrounding DRBD which is now 
> horribly out of date).
>
> I would change this:
>  cs_order { ''jboss_ip_before_jboss'':
>     first   => ''applayer_ip'',
>     second  => ''applayer_jboss'',
>     require => Cs_colocation[''jboss_with_ip''],
>     }
> to
>  cs_order { ''jboss_ip_before_jboss'':
>     first   => ''applayer_ip'',
>     second  => ''applayer_jboss'',
>     score => ''INFINITY'',
>     require =>
Cs_primitive[''applayer_ip'',''applayer_jboss''],
>     }
> and 
> cs_colocation { ''jboss_with_ip'':
>       primitives          => [ ''applayer_ip'',
''applayer_jboss'' ],
>       require =>
Cs_primitive[''applayer_ip'',''applayer_jboss''],
>     }
>
> as autorequires were very, very dicey when I mucked with the module. 
>  Ignore this if they work now.
>
> Also - make sure you''re not setting up the IP or the service
anywhere else
> as corosync should be managing it.
>
> Outside of puppet, are you sure IPaddr2 is the right primitive_class?  Can 
> we get the output of a crm configure show, and if it''s missing the
> primitive if it takes:
> crm configure primitive applayer_ip ocf:heartbeat:IPaddr2 params 
> ip="whatever" cidr_netmask="whatever"
>
> because I suspect that''s the problem.
>
>
-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/puppet-users/-/-caTbsSwxhMJ.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.