moby/libnetwork/networkdb
Flavio Crisciani d6440c9139 optimize the rebroadcast for failure case
Before when a node was failing, all the nodes would bump the lamport time of all their
entries. This means that if a node flap, there will be a storm of update of all the entries.
This commit on the base of the previous logic guarantees that only the node that joins back
will readvertise its own entries, the other nodes won't need to advertise again.

Signed-off-by: Flavio Crisciani <flavio.crisciani@docker.com>
2017-08-01 14:08:54 -07:00
..
broadcast.go 2x faster to converge 2017-08-01 13:47:18 -07:00
cluster.go optimize the rebroadcast for failure case 2017-08-01 14:08:54 -07:00
delegate.go fix join/leave 2017-08-01 14:08:45 -07:00
event_delegate.go optimize the rebroadcast for failure case 2017-08-01 14:08:54 -07:00
message.go NetworkDB allow setting PacketSize 2017-07-26 13:44:33 -07:00
networkdb.go optimize the rebroadcast for failure case 2017-08-01 14:08:54 -07:00
networkdb.pb.go Recover from transient gossip failures 2016-09-19 15:58:14 -07:00
networkdb.proto Recover from transient gossip failures 2016-09-19 15:58:14 -07:00
networkdb_test.go NetworkDB allow setting PacketSize 2017-07-26 13:44:33 -07:00
networkdbdiagnose.go NetworkDB testing infra 2017-07-27 08:50:43 -07:00
watch.go Fix leak of handleTableEvents 2017-05-31 11:04:19 -07:00