httpd restarts on every chef run about chef-logstash HOT 3 CLOSED

could it be related to a json hash serializing in a different order?

from chef-logstash.

I had a similar problem with my logstash cookbook - every run would restart logstash on my servers. Turned out the search for ElasticSearch nodes was returning them in a random order. So when I grabbed the first one from the list it was often not the one the config pointed to currently. And as a result, it would re-write the config file and restart Logstash.

Is it possible the same thing is happening here with your Kibana config? To fix my issue I just put a sort in after I got the list to ensure I was consistently getting the same server if the IPs were the same.

    es_results = search(:node, "roles:#{node['logstash']['elasticsearch_role']} AND chef_environment:#{node.chef_environment}")
    es_results.sort! { |a,b| a.name.downcase <=> b.name.downcase }
    es_server_ip = es_results[0]['ec2']['public_hostname']

It doesn't do any load balancing, all servers will get the same ES node. But it was a simple fix to stop the restarts (especially with logstash jar startup time). Eventually I'll probably front my ES cluster with an a load balancer and that should fix it for good.

from chef-logstash.

closing as we've pulled out kibana.

from chef-logstash.