Merge pull request #2370 from hashicorp/f-gc

Allow specification of eval/job gc threshold

Author: dadgar

command/agent/agent.go

@@ -165,6 +165,20 @@ func convertServerConfig(agentConfig *Config, logOutput io.Writer) (*nomad.Confi
 		}
 		conf.NodeGCThreshold = dur
 	}
+	if gcThreshold := agentConfig.Server.JobGCThreshold; gcThreshold != "" {
+		dur, err := time.ParseDuration(gcThreshold)
+		if err != nil {
+			return nil, err
+		}
+		conf.JobGCThreshold = dur
+	}
+	if gcThreshold := agentConfig.Server.EvalGCThreshold; gcThreshold != "" {
+		dur, err := time.ParseDuration(gcThreshold)
+		if err != nil {
+			return nil, err
+		}
+		conf.EvalGCThreshold = dur
+	}
 
 	if heartbeatGrace := agentConfig.Server.HeartbeatGrace; heartbeatGrace != "" {
 		dur, err := time.ParseDuration(heartbeatGrace)

command/agent/config-test-fixtures/basic.hcl

@@ -65,6 +65,8 @@ server {
 	num_schedulers = 2
 	enabled_schedulers = ["test"]
 	node_gc_threshold = "12h"
+	job_gc_threshold = "12h"
+	eval_gc_threshold = "12h"
 	heartbeat_grace   = "30s"
 	retry_join = [ "1.1.1.1", "2.2.2.2" ]
 	start_join = [ "1.1.1.1", "2.2.2.2" ]

command/agent/config.go

@@ -238,8 +238,20 @@ type ServerConfig struct {
 	EnabledSchedulers []string `mapstructure:"enabled_schedulers"`
 
 	// NodeGCThreshold controls how "old" a node must be to be collected by GC.
+	// Age is not the only requirement for a node to be GCed but the threshold
+	// can be used to filter by age.
 	NodeGCThreshold string `mapstructure:"node_gc_threshold"`
 
+	// JobGCThreshold controls how "old" a job must be to be collected by GC.
+	// Age is not the only requirement for a Job to be GCed but the threshold
+	// can be used to filter by age.
+	JobGCThreshold string `mapstructure:"job_gc_threshold"`
+
+	// EvalGCThreshold controls how "old" an eval must be to be collected by GC.
+	// Age is not the only requirement for a eval to be GCed but the threshold
+	// can be used to filter by age.
+	EvalGCThreshold string `mapstructure:"eval_gc_threshold"`
+
 	// HeartbeatGrace is the grace period beyond the TTL to account for network,
 	// processing delays and clock skew before marking a node as "down".
 	HeartbeatGrace string `mapstructure:"heartbeat_grace"`
@@ -834,6 +846,12 @@ func (a *ServerConfig) Merge(b *ServerConfig) *ServerConfig {
 	if b.NodeGCThreshold != "" {
 		result.NodeGCThreshold = b.NodeGCThreshold
 	}
+	if b.JobGCThreshold != "" {
+		result.JobGCThreshold = b.JobGCThreshold
+	}
+	if b.EvalGCThreshold != "" {
+		result.EvalGCThreshold = b.EvalGCThreshold
+	}
 	if b.HeartbeatGrace != "" {
 		result.HeartbeatGrace = b.HeartbeatGrace
 	}

command/agent/config_parse.go

@@ -498,6 +498,8 @@ func parseServer(result **ServerConfig, list *ast.ObjectList) error {
 		"num_schedulers",
 		"enabled_schedulers",
 		"node_gc_threshold",
+		"eval_gc_threshold",
+		"job_gc_threshold",
 		"heartbeat_grace",
 		"start_join",
 		"retry_join",

command/agent/config_parse_test.go

@@ -82,6 +82,8 @@ func TestConfig_Parse(t *testing.T) {
 					NumSchedulers:     2,
 					EnabledSchedulers: []string{"test"},
 					NodeGCThreshold:   "12h",
+					EvalGCThreshold:   "12h",
+					JobGCThreshold:    "12h",
 					HeartbeatGrace:    "30s",
 					RetryJoin:         []string{"1.1.1.1", "2.2.2.2"},
 					StartJoin:         []string{"1.1.1.1", "2.2.2.2"},

website/source/docs/agent/configuration/server.html.md

@@ -68,6 +68,14 @@ server {
   terminal state before it is garbage collected and purged from the system. This
   is specified using a label suffix like "30s" or "1h".
 
+- `job_gc_threshold` `(string: "4h")` - Specifies the minimum time a job must be
+  in the terminal state before it is eligible for garbage collection. This is
+  specified using a label suffix like "30s" or "1h".
+
+- `eval_gc_threshold` `(string: "1h")` - Specifies the minimum time an
+  evaluation must be in the terminal state before it is eligible for garbage
+  collection. This is specified using a label suffix like "30s" or "1h".
+
 - `num_schedulers` `(int: [num-cores])` - Specifies the number of parallel
   scheduler threads to run. This can be as many as one per core, or `0` to
   disallow this server from making any scheduling decisions. This defaults to