raboof/pekko
1
0
Fork 0
Code Issues Pull requests Projects Releases 6 Packages Wiki Activity Actions 10
12931 commits 1 branch 31 tags 268 MiB
Commit graph

992 commits

Author SHA1 Message Date
Patrik Nordwall
196a141976 FIXME in cluster, see #3192 2013-05-28 09:02:03 +02:00
Patrik Nordwall
c05262f406 Change default tick-duration to 10ms, see #3387 2013-05-27 09:20:25 +02:00
Patrik Nordwall
28d1b1f187 Merge pull request #1480 from akka/wip-3388-HeartbeatReq-patriknw 
Start heartbeatSender after Welcome, see #3388
 2013-05-27 00:17:43 -07:00
Patrik Nordwall
3736efb79a Merge pull request #1472 from akka/wip-3225-cluster-infolog-patriknw 
Config of cluster info logging, see #3225
 2013-05-27 00:13:08 -07:00
Patrik Nordwall
8232ebc5eb Merge pull request #1471 from akka/wip-3252-MemberRemoved-patriknw 
Add previousStatus in MemberRemoved, see #3252
 2013-05-27 00:12:48 -07:00
Patrik Nordwall
ec1626b746 Start heartbeatSender after Welcome, see #3388 
* Otherwise, if the Welcome message is lost, other nodes
  in the cluster will send HeartbeatRequest and it will start
  sending heartbeats without being a real member and the lost Welcome
  is not detected by the other members in the cluster
 2013-05-24 15:38:28 +02:00
Patrik Nordwall
18a3b3facf Config of cluster info logging, see #3225 2013-05-23 13:36:35 +02:00
Patrik Nordwall
ee6e80d31a Add previousStatus in MemberRemoved, see #3252 2013-05-23 11:09:32 +02:00
Patrik Nordwall
f29447e0b2 Increase timeout in MetricsCollectorSpec, see #3358 2013-05-23 08:54:21 +02:00
drewhk
336fb1b180 Merge pull request #1450 from drewhk/wip-eventstream-for-fd-drewhk 
EventStream is now passed to failure detectors
 2013-05-17 09:40:12 -07:00
Endre Sándor Varga
fd4bc09035 EventStream is now passed to failure detectors 
- Also, dynamic loading is now centralized (DRY)
 2013-05-17 16:35:27 +02:00
Patrik Nordwall
8f04b53ac7 Merge pull request #1443 from akka/wip-3359-auto-join-patriknw 
Remove auto-join config, derive from seed-nodes, see #3359
 2013-05-17 04:57:07 -07:00
Patrik Nordwall
ad1eaa6d4a Remove auto-join config, derive from seed-nodes, see #3359 2013-05-17 13:54:51 +02:00
Patrik Nordwall
a0a0f39613 Hardening of cluster member leaving path, see #3309 
* Removed leader commands for Shutdown and Exit
* Member shutdown itself  when it sees itself as Exiting
* Singleton cluster with status Exiting will shutdown itself,
  in case the Exiting gossip never arrives
* Exiting member not part convergence check
* Exiting member is removed by leader (on convergence) when the
  exiting member is in the unreachable set, i.e. sucessfully shutdown
* Reverted the change made for #3266, i.e. Exiting is
  detected as unreachable again.
* Adjust ClusterSingletonManager to new Exiting behaviour
* Fix bug in HeartbeatSender, which caused it to continue to
  send heartbeats to removed nodes, instead of rebalancing
* Refactoring of leaderActions method
* Leaving section in docs
 2013-05-17 11:39:49 +02:00
Patrik Nordwall
713e323265 Compression of large cluster messages, see #3355 2013-05-14 21:43:12 +02:00
Patrik Nordwall
d0ed7385b2 Merge pull request #1420 from akka/wip-3296-pub-metrics-patriknw 
Publish cluster metrics periodically, see #3296
 2013-05-13 06:02:32 -07:00
Patrik Nordwall
9ec0b7ca71 Better test timeouts in UnreachableNodeJoinsAgainSpec, see #3285 2013-05-09 21:27:23 +02:00
Patrik Nordwall
ace0c53a6f Publish cluster metrics periodically, see #3296 
* instead of when it receives metrics gossip and when it collects metrics
 2013-05-09 20:54:46 +02:00
Patrik Nordwall
236330f3c6 Merge pull request #1401 from akka/wip-3265-heartbeating-race2-patriknw 
Changed design of RemoteWatcher due to cleanup race, see #3265
 2013-05-07 04:27:26 -07:00
Björn Antonsson
e00ab533bb Wait on shutdown of extra actor systems in tests. See #3217 2013-05-07 11:02:03 +02:00
Patrik Nordwall
7628889b43 Changed design of RemoteWatcher due to cleanup race, see #3265 
* The problem was a race caused by HeartbeatReq sent out, and
  the watchee terminated immediately. That caused the RemoteWatcher
  peers watching each other without any other watch registered.
  It is racy.
* Instead of one-way heartbeats from the side beeing watched I
  changed to ping-pong style. That makes the problem go away
  and simplifies a lot of things in RemoteWatcher.
 2013-05-04 17:35:12 +02:00
Roland
b3db19ee05 Merge branch 'wip-3281-NullMessage-∂π' 2013-05-03 19:40:36 +02:00
Patrik Nordwall
b8b65c9153 Cluster member age, and usage in singleton, see #3195 
* Assign internal upNumber when member is moved to Up
* Public API Member.isOlder
* Change cluster singleton to use oldest member instead of leader
* Update samples and docs
 2013-05-03 13:38:35 +02:00
Björn Antonsson
459e3e3720 Merge pull request #1378 from akka/wip-3273-enforce-mailbox-types-on-system-actors-ban 
Enforce mailbox types on System actors. #3273
 2013-05-03 04:04:02 -07:00
Björn Antonsson
539df2e98a Enforce mailbox types on System actors. See #3273 2013-05-03 11:05:32 +02:00
Patrik Nordwall
a9394d9f71 Remove duplicates in roleMapping, see #3270 2013-05-03 10:07:51 +02:00
Patrik Nordwall
e2b0fdeff9 Move barriers of conditional tests in StressSpec, see #2787 2013-05-02 20:23:14 +02:00
Patrik Nordwall
6635ac4032 Reduce amount of gossip data transferred in idle cluster, see #3279 
* When seen same the gossip chat is initated with GossipStatus
  message containing the vclock only
* Remove conversation flag in GossipEnvelope
* Ordinary tell instead of actorSelection when replying
 2013-05-02 19:17:09 +02:00
Patrik Nordwall
671ebf8909 Additional tests of ClusterMessageSerializer 2013-05-02 19:17:08 +02:00
Patrik Nordwall
293c97c71d Quick fix for unreachable exiting, see #3266 2013-05-02 19:17:08 +02:00
Patrik Nordwall
33a8808a6d Enable usage of MultiJvm nrOfNodes in cluster StressSpec, see #2787 
* Adjustments to StressSpec for testing large clusters
* Performance improvement of mute deadLetters
 2013-05-02 19:17:08 +02:00
Roland
738796c625 remove NullMessage, see #3281 2013-05-02 18:48:36 +02:00
Roland Kuhn
927a00ee5a Merge pull request #1382 from drexin/wip-drexin-remove-deprecations 
Removed lots of deprecation warnings caused by usages of deprecated Props creators
 2013-04-30 05:41:11 -07:00
Patrik Nordwall
4900f562b5 Merge pull request #1374 from akka/wip-3265-watcher-crash-patriknw 
Stop heartbeating when watching node crash, see #3265
 2013-04-28 13:16:51 -07:00
dario.rexin
3e8597d94b more deprecation warnings removed 2013-04-26 13:54:10 +02:00
Björn Antonsson
5b0c575146 Merge pull request #1364 from akka/wip-3136-make-fromconfig-supervisorstrategy-configurable-ban 
Allow a FromConfig router to be partly configured from code #3136
 2013-04-26 00:22:05 -07:00
Patrik Nordwall
551e2d1321 Stop heartbeating when watching node crash, see #3265 2013-04-25 21:25:46 +02:00
Patrik Nordwall
b62f5f46f5 Merge pull request #1361 from akka/wip-3247-UnreachableNodeJoinsAgainSpec-patriknw 
Solve UnreachableNodeJoinsAgainSpec problem, see #3247
 2013-04-25 05:32:08 -07:00
Patrik Nordwall
19283207ff Additional coordination of shutdown in ClusterDeathWatchSpec, see #3255 
* Added testconductor shutdown of actor system
* Renamed previous shutdown to exit
 2013-04-24 11:58:07 +02:00
Patrik Nordwall
2ebb2a0b9c Solve UnreachableNodeJoinsAgainSpec problem, see #3247 
* UnreachableNodeJoinsAgain failed because of gated connection
* Removed default test value of retry-gate-closed-for, instead
  default from reference.conf is used, i.e. 0s
* deadLetters logging love
 2013-04-23 15:43:10 +02:00
Patrik Nordwall
428e71690f Coordinate shutdown of ClusterDeathWatchSpec with messages, see #3255 
* Added MultiNodeClusterSpec.EndActor for this purpose
* Changed UnreachableNodeJoinsAgainSpec to use the same
 2013-04-23 11:56:12 +02:00
Björn Antonsson
6aa7dc7f97 Allow a FromConfig router to be partly configured from code. See #3136 2013-04-23 11:36:10 +02:00
drewhk
a694995617 Merge pull request #1352 from drewhk/wip-2594-quarantining-drewhk 
Hooked in quarantining
 2013-04-19 06:02:08 -07:00
Endre Sándor Varga
fa5ab15450 Ignoring test that relies on actorFor and remote watch 2013-04-19 14:51:21 +02:00
Björn Antonsson
c3eed374f1 Allow different types of mailboxes on the same dispatcher. See #2687 2013-04-19 13:42:58 +02:00
Patrik Nordwall
d04d44b184 Merge pull request #1349 from akka/wip-2594-quarantine-removed-member-patriknw 
Only quarantine removed member that was unreachable, see #2594
 2013-04-19 03:02:32 -07:00
Patrik Nordwall
49744e0b0f Only quarantine removed member that was unreachable, see #2594 
* For graceful leaving and remove it should still be possible to
  communicate with the node after cluster removal.
* Otherwise the hand over in cluster singleton would break, for
  example.
* Also, skip selfAddress to avoid generation of AddressTerminated
  for the own node when removed from cluster.
 2013-04-19 08:52:27 +02:00
Patrik Nordwall
a3df775931 Change member prio for Joining/Up, see #3239 
* Sometimes caused duplicate MemberUp events, after conflicting gossips
 2013-04-19 07:54:23 +02:00
Patrik Nordwall
c7db89e9ab Merge pull request #1344 from akka/wip-quarantining-in-ClusterRemoteWatcher-patriknw 
Quarantine from ClusterRemoteWatcher also, see #2993
 2013-04-18 13:05:06 -07:00
drewhk
98d4bcc9dd Merge pull request #1338 from drewhk/wip-1478-sysmsg-guaranteed-delivery-drewhk 
Guaranteed delivery of system messages (and now praying...)
 2013-04-18 12:38:28 -07:00