Fix name and use of CORE_CLUSTER_RECOVER_TIMEOUT_SEC

datarhei · Jul 9, 2024 · 7879480 · 7879480
1 parent e306d20
commit 7879480
Show file tree

Hide file tree

Showing 3 changed files with 5 additions and 2 deletions.
diff --git a/cluster/cluster.go b/cluster/cluster.go
@@ -55,6 +55,7 @@ type Cluster interface {
 	Leave(origin, id string) error              // gracefully remove a node from the cluster
 	TransferLeadership(origin, id string) error // transfer leadership to another node
 	Snapshot(origin string) (io.ReadCloser, error)
+	IsRaftLeader() bool
 	HasRaftLeader() bool
 
 	ProcessAdd(origin string, config *app.Config) error

diff --git a/cluster/follower.go b/cluster/follower.go
@@ -33,7 +33,9 @@ func (c *cluster) establishFollowership(ctx context.Context) {
 	ctx, cancel := context.WithCancel(ctx)
 	c.cancelFollowerShip = cancel
 
-	go c.recoverCluster(ctx, c.syncInterval)
+	if c.recoverTimeout > 0 {
+		go c.recoverCluster(ctx, c.syncInterval)
+	}
 }
 
 func (c *cluster) revokeFollowership() {

diff --git a/config/config.go b/config/config.go
@@ -299,7 +299,7 @@ func (d *Config) init() {
 	d.vars.Register(value.NewInt64(&d.Cluster.SyncInterval, 5), "cluster.sync_interval_sec", "CORE_CLUSTER_SYNC_INTERVAL_SEC", nil, "Interval between aligning the process in the cluster DB with the processes on the nodes", true, false)
 	d.vars.Register(value.NewInt64(&d.Cluster.NodeRecoverTimeout, 120), "cluster.node_recover_timeout_sec", "CORE_CLUSTER_NODE_RECOVER_TIMEOUT_SEC", nil, "Timeout for a node to recover before rebalancing the processes", true, false)
 	d.vars.Register(value.NewInt64(&d.Cluster.EmergencyLeaderTimeout, 10), "cluster.emergency_leader_timeout_sec", "CORE_CLUSTER_EMERGENCY_LEADER_TIMEOUT_SEC", nil, "Timeout for establishing the emergency leadership after lost contact to raft leader", true, false)
-	d.vars.Register(value.NewInt64(&d.Cluster.RecoverTimeout, 0), "cluster.recover_timeout_sec", "CORE_CLUSTER_NODE_RECOVER_TIMEOUT_SECONDS", nil, "Timeout for recovering the cluster if no leader can be voted", false, false)
+	d.vars.Register(value.NewInt64(&d.Cluster.RecoverTimeout, 0), "cluster.recover_timeout_sec", "CORE_CLUSTER_RECOVER_TIMEOUT_SEC", nil, "Timeout for recovering the cluster if no raft leader can be elected", false, false)
 	d.vars.Register(value.NewBool(&d.Cluster.Debug.DisableFFmpegCheck, false), "cluster.debug.disable_ffmpeg_check", "CORE_CLUSTER_DEBUG_DISABLE_FFMPEG_CHECK", nil, "Disable checking for identical FFmpeg versions on all nodes", false, false)
 }