i have a setup as below,
Rabbitmq central cluster servers with federation:
Canada location cluster servers which used to connect all canada location servers(sub collective name as canada_mcollective):
Texas location cluster servers which used to connect all canada location servers(sub collective name as texas_mcollective):
configured federation upstream on centram cluster as below:
virtual host : /mcollective
ack mode : on-confirm
expirt : 36000000
federation upstream policy:
virtual host : /mcollective
Pattern : ^.*
apply to : all
definition : federation-upstream-set: all
Priority : 0
I created exhanges as below,
mcollective_broadcast -> topic -> durable
mcollective_directed -> direct -> durable
mcollective_reply -> direct -> durable
canada_mcollective_broadcast -> topic -> durable
canada_mcollective_directed -> direct -> durable
texas_mcollective_broadcast -> topic -> durable
texas_mcollective_directed -> direct -> durable
so when do mco ping am not getting response from all the connected servers and sometimes am getting response from some server. Mean not getting stable response.
I checked on logs and found that exchanges are fordwarded to each location federatiion cluster.
But if do mco shell for each server name with their identity then am getting proper response. I think am missing something here to work with geographical federation configuration.
also i have added the parameter "plugin.rabbitmq.use_reply_exchange = true" on both server and client cfg file.
I checked the mcollective log on each servers and looks like all the servers got the message and responded accordingly and consumer is not receiving it properly.
So I suspect since its cluster it might conflicting between them so I tried configuring HA Proxy as well which also doesn't help.
Any help please..