Merge pull request #3725 from weaveworks/retry-pending-claims

Retry IP allocator claims

ipam/allocator.go

@@ -275,7 +275,7 @@ func (alloc *Allocator) spaceRequestDenied(sender mesh.PeerName, r address.Range
 	for i := 0; i < len(alloc.pendingClaims); {
 		claim := alloc.pendingClaims[i].(*claim)
 		if r.Contains(claim.cidr.Addr) {
-			claim.deniedBy(alloc, sender)
+			claim.deniedBy(alloc, sender, nil)
 			alloc.pendingClaims = append(alloc.pendingClaims[:i], alloc.pendingClaims[i+1:]...)
 			continue
 		}

ipam/claim.go

@@ -12,6 +12,7 @@ import (
 
 type claim struct {
 	resultChan       chan<- error
+	tryCount         int
 	ident            string       // a container ID, something like "weave:expose", or api.NoContainerID
 	cidr             address.CIDR // single address being claimed
 	isContainer      bool         // true if ident is a container ID
@@ -19,6 +20,8 @@ type claim struct {
 	hasBeenCancelled func() bool
 }
 
+const maxTryCount = 5
+
 // Send an error (or nil for success) back to caller listening on resultChan
 func (c *claim) sendResult(result error) {
 	// Make sure we only send a result once, since listener stops listening after that
@@ -40,6 +43,8 @@ func (c *claim) Try(alloc *Allocator) bool {
 		return true
 	}
 
+	c.tryCount++
+
 	addOwned := func() {
 		if c.ident == api.NoContainerID {
 			alloc.addOwned(c.cidr.Addr.String(), c.cidr, c.isContainer)
@@ -90,8 +95,10 @@ func (c *claim) Try(alloc *Allocator) bool {
 			if c.noErrorOnUnknown {
 				alloc.infof("Claim %s for %s: %s; will try later.", c.cidr, c.ident, err)
 				c.sendResult(nil)
-			} else { // just tell the user they can't do this.
-				c.deniedBy(alloc, owner)
+			} else if c.tryCount > maxTryCount {
+				// give up, tell the user they can't do this
+				c.deniedBy(alloc, owner, err)
+				return true
 			}
 		}
 		return false
@@ -128,12 +135,16 @@ func (c *claim) Try(alloc *Allocator) bool {
 	return true
 }
 
-func (c *claim) deniedBy(alloc *Allocator, owner mesh.PeerName) {
+func (c *claim) deniedBy(alloc *Allocator, owner mesh.PeerName, err error) {
 	name, found := alloc.nicknames[owner]
 	if found {
 		name = " (" + name + ")"
 	}
-	c.sendResult(fmt.Errorf("address %s is owned by other peer %s%s", c.cidr.String(), owner, name))
+	reason := ""
+	if err != nil {
+		reason = fmt.Sprintf(" - %s", err)
+	}
+	c.sendResult(fmt.Errorf("address %s is owned by other peer %s%s%s", c.cidr.String(), owner, name, reason))
 }
 
 func (c *claim) Cancel() {