Use Chef search (role:cassandra) to look up seed nodes

attributes/default.rb

@@ -31,6 +31,14 @@
 default['cassandra']['log_dir'] = '/var/log/cassandra'
 default['cassandra']['rootlogger'] = 'INFO,stdout,R'
 
+# Seed node discovery
+default['cassandra']['seeds'] = node['ipaddress']
+
+default['cassandra']['seed_discovery']['use_chef_search'] = false
+default['cassandra']['seed_discovery']['count'] = 3
+default['cassandra']['seed_discovery']['search_role'] = 'cassandra-seed'
+default['cassandra']['seed_discovery']['search_query'] = nil
+
 default['cassandra']['jbod']['slices'] = nil
 default['cassandra']['jbod']['dir_name_prefix'] = 'data'
 
@@ -129,7 +137,6 @@
 default['cassandra']['xss'] = '256k'
 default['cassandra']['vnodes'] = true
 default['cassandra']['num_tokens'] = 256
-default['cassandra']['seeds'] = node['ipaddress']
 default['cassandra']['enable_assertions'] = true
 default['cassandra']['internode_compression'] = 'all' # all, dc, none
 default['cassandra']['jmx_server_hostname'] = false

libraries/config_helpers.rb

@@ -18,3 +18,36 @@ def cassandra_bool_config(config_val)
 def hash_to_yaml_string(hash)
   hash.to_hash.to_yaml
 end
+
+def discover_seed_nodes
+  # use chef search for seed nodes
+  if node['cassandra']['seed_discovery']['use_chef_search']
+    if Chef::Config[:solo]
+      Chef::Log.warn("Chef Solo does not support search, provide the seed nodes via node attribute node['cassandra']['seeds']")
+      node['ipaddress']
+    else
+      Chef::Log.info("Cassandra seed discovery using Chef search is enabled")
+      q = if search_query = node['cassandra']['seed_discovery']['search_query']
+            search_query
+          else
+            "chef_environment:#{node.chef_environment} AND role:#{node['cassandra']['seed_discovery']['search_role']} AND cassandra_cluster_name:#{node['cassandra']['cluster_name']}"
+          end
+      Chef::Log.info("Will discover Cassandra seeds using query '#{q}'")
+      xs = search(:node, q).map(&:ipaddress).sort.uniq
+      Chef::Log.debug("Discovered #{xs.size} Cassandra seeds using query '#{q}'")
+
+      if xs.empty?
+        node['ipaddress']
+      else
+        xs.take(node['cassandra']['seed_discovery']['count']).join(',')
+      end
+    end
+  else
+    # user defined seed nodes
+    if node['cassandra']['seeds'].is_a?(Array)
+      node['cassandra']['seeds'].join(',')
+    else
+      node['cassandra']['seeds']
+    end
+  end
+end

recipes/default.rb

@@ -35,4 +35,6 @@
   node.default['cassandra']['jamm']['sha256sum'] = '79d44f1b911a603f0a249aa59ad6ea22aac9c9b211719e86f357646cdf361a42'
 end
 
+node.default['cassandra']['seeds'] = discover_seed_nodes
+
 include_recipe "cassandra-dse::#{node['cassandra']['install_method']}"

templates/default/cassandra.yaml.erb

@@ -242,7 +242,7 @@ seed_provider:
           # seeds is actually a comma-delimited list of addresses.
           # Ex: "<ip1>,<ip2>,<ip3>"
           #- seeds: "127.0.0.1"
-          - seeds: "<%= if node['cassandra']['seeds'].kind_of?(Array) then node['cassandra']['seeds'].join(",") else node['cassandra']['seeds'] end %>"
+          - seeds: "<%= node['cassandra']['seeds'] %>"
 
 # emergency pressure valve: each time heap usage after a full (CMS)
 # garbage collection is above this fraction of the max, Cassandra will