You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have been trying to get the aws auto peer discovery working with my auto scaling group using this cookbook, but the problem I have ran into is the config file formats. This recipe sources the rabbitmq.config.erb template and will always create that in the /etc/rabbitmq directory.
In order for me to get the auto discovery working manually, I had to delete the .config file so that the service would pickup my rabbitmq.conf file instead. This is because I am using the new sysctl format for the config file (Key = Value style).
Any plans to update this and add an attribute to not source the .config (erlang style config) in the default recipe when a user is wanting to use the new .conf (sysctl) style?
I was going to fork this and try myself, but figured I'd ask first. I could just be doing something completely wrong as well. This is my first stab at RMQ, let alone auto clustering :)
Thanks all in advance.
The text was updated successfully, but these errors were encountered:
New style config format is something I've been working on for a while. I make no ETA promises but there seems to be nothing specific to peer discovery or AWS about that: you just have to wait till 6.0. It's not obvious how to progress to the new format without introducing too many breaking changes at once, that's the biggest problem.
Peer discovery plugins (just like everything else) can use the classic config format since the new one is translated to the old one with extra validations anyway. If you have questions about that, please post them to the mailing list (rabbitmq-users).
Thank you @michaelklishin - I have a workaround. I'm going to only run the recipe once on the nodes as they are provisioned. I can add custom code in my wrapper cookbook to facilitate the changes I need for the new version of Rabbit.
Thanks for the prompt reply and explanation! Cheers.
Also worth mentioning that this cookbook already supports arbitrary configuration, although it's pretty brittle. So if you are up for maintaining a wrapper, I can easily see how that'd be the least painful option for now.
You can watch #497 for the general 3.7.x config files/format support.
Cookbook version used
5.6.2
Operating system, version, and patch level
Red Hat Enterprise Linux Server release 7.5 (Maipo) - all up to date
Node attributes
node.default['rabbitmq']['version'] = '3.7.7'
node.default['rabbitmq']['clustering']['enable'] = true
node.default['rabbitmq']['clustering']['use_auto_clustering'] = true
node.default['pilot_rabbitmq']['key_id'] = 'XXXXXXXX'
node.default['pilot_rabbitmq']['secret_key'] = 'XXXXXXXX'
node.default['rabbitmq']['erlang_cookie'] = 'XXXXXXXXX'
node.default['erlang']['install_method'] = "esl"
node.default['erlang']['esl']['version'] = '19.3'
RabbitMQ version provisioned
3.7.7
Erlang version provisioned
19.3
I have been trying to get the aws auto peer discovery working with my auto scaling group using this cookbook, but the problem I have ran into is the config file formats. This recipe sources the rabbitmq.config.erb template and will always create that in the /etc/rabbitmq directory.
In order for me to get the auto discovery working manually, I had to delete the .config file so that the service would pickup my rabbitmq.conf file instead. This is because I am using the new sysctl format for the config file (Key = Value style).
Any plans to update this and add an attribute to not source the .config (erlang style config) in the default recipe when a user is wanting to use the new .conf (sysctl) style?
I was going to fork this and try myself, but figured I'd ask first. I could just be doing something completely wrong as well. This is my first stab at RMQ, let alone auto clustering :)
Thanks all in advance.
The text was updated successfully, but these errors were encountered: