* Rename config akka.event-handlers to akka.loggers
* Rename config akka.event-handler-startup-timeout to
akka.logger-startup-timeout
* Rename JulEventHandler to JavaLogger
* Rename Slf4jEventHandler to Slf4jLogger
* Change all places in tests and docs
* Deprecation, old still works, but with warnings
* Migration guide
* Test for the deprecated event-handler config
* Failure detector was previously copied with refactoring to
akka-remote and this refactoring makes use of that and removes
the failure detector in akka-cluster
* Adjustments to reference.conf
* Refactoring of FailureDetectorPuppet
- New DeadLetter class for handling remoting specific envelopes
- Fixed error handling of name lookups
- Name lookup is now handled via futures (future refactor opportunity)
* When a def starts with if and is not a oneliner the if
should be on a new line.
* The reason is that it might be easy to miss the if when
reading the code.
* Subscribe to InstantMemberEvent and start heartbeating when
InstantMemberUp. Same for metrics.
* HeartbeatNodeRing data structure for bidirectional mapping of
heartbeat sender and receiver. Not using ConsistentHash anymore.
Node addresses are hashed to ensure that neighbors are spread out.
* HeartbeatRequest when receiver detects that it has not received
expected heartbeats.
* New test InitialHeartbeatSpec that simulates the problem
* Add/remove some related conf properties
* Add some more logging to be able to diagnose eventual problems
* Explicit config of nr-of-end-heartbeats
* akka.cluster.StressSpec
* Configurable number of nodes and duration for each step
* Report metrics and phi periodically to see progress
* Configurable payload size
* Test of various join and remove scenarios
* Test of watch
* Exercise supervision
* Report cluster stats
* Test with many actors in tree structure
Apart from the test this commit also solves some issues:
* Avoid adding back members when downed in ClusterHeartbeatSender
* Avoid duplicate close of ClusterReadView
* Add back the publish of AddressTerminated when MemberDowned/Removed
it was lost in merge of "publish on convergence", see #2779
* Instead of when unreachable
* Note that ClusterRouterConfig is not changed, i.e. routees will be removed
when unreachable
* Routers that are not wrapped by ClusterRouterConfig will watch as usual, i.e.
remove routees when Terminated, i.e. node down
* Renamed isRunning to isTerminated (with negation of course)
* Removed Running from JMX API, since the mbean is deregistered anyway
* Cleanup isAvailable, isUnavailbe
* Misc minor
* MBeanSpec
* Added Members and Unreachable to JMX API
* Removed Convergence from JMX API, because it will
not be exposed when ticket #2692 is merged
* Updated documentation and akka-cluster script
It's a glitch in how ClusterReadView (used in tests) is updated.
First we do awaitUpConvergence, which checks readView.members and
readView.convergence. Then we do assertLeader, which checks
readView.leader. The problem is that readView.leader might not
have been updated yet (if there already was convergence).
Solution is to await the expected leader in awaitUpConvergence,
so that readView is in a consistent state after awaitUpConvergence.
We will change the semantics in ticket #2692, but this change will
be needed and should work with that as well.
These tests use the throttling in the experimental test conductor which relies
on the fact that the same connection is used for both inbound and outbound
traffic. This is not always the case when starting multiple cluster nodes
at the same time.
* case object and case class for MixMetricsSelector
* Rename decay-half-life-duration to moving-average-half-life
* Clarification of decay-half-life-duration and collect-interval
* Removed Fields, Java compatibility issue
* Adapt for-yield variables
* Comment metrics collector constructor that takes system param
* Don't copy EWMA if not needed
* LogOf2 constant 0.69315
* Don't use mapValues
* Remove RichInt conversion
* sigar version replace tag in docs
* createDeployer factory method to make it possible to override
deployer in subclass
* Improve readability of MetricsListener (in sample)
* Better startup of factorial sample (no sleep)
* Many minor enhancements and cleanups
* Refactoring of standard metrics extractors and data structures
* Removed optional value in Metric, simplified a lot
* Configuration of EWMA by using half-life duration
* Renamed DataStream to EWMA
* Incorporate review feedback
* Use binarySearch for selecting weighted routees
* More metrics selectors for the router
* Removed network metrics, since not supported on linux
* Configuration of router
* Rename to AdaptiveLoadBalancingRouter
* Remove total cores metrics, since it's the same as jmx getAvailableProcessors,
tested on intel 24 core server and amd 48 core server, and MBP
* API cleanup
* Java API additions
* Documentation of metrics and AdaptiveLoadBalancingRouter
* New cluster sample to illustrate metrics in the documentation,
and play around with (factorial)
* MetricsSelector, calculate capacity, weights and allocate weighted
routee refs
* ClusterLoadBalancingRouterSpec
* Optional heap max
* Constants for the metric fields
* Refactoring of Metric and decay
* Rewrite of DataStreamSpec
* Correction of EWMA and removal of BigInt, BigDecimal
* Separation of MetricsCollector into trait and two classes,
SigarMetricsCollector and JmxMetricsCollector
* This will reduce cost when sigar is not installed, such as
avoiding throwing and catching exc for every call
* Improved error handling for loading sigar
* Made MetricsCollector implementation configurable
* Tested with sigar
* The problem was that we didn't wait for the testconductor.shutdown Future
to complete and therefore barriers could be triggered in unexpected order.
The reason why we didn't await, was that during shutdown the Future was
completed with client disconnected failure. I have fixed that and added
await to all shutdowns.
* Previously heartbeat messages was sent to all other members, i.e.
each member was monitored by all other members in the cluster.
* This was the number one know scalability bottleneck, due to the
number of interconnections.
* Limit sending of heartbeats to a few (5) members. Select and
re-balance with consistent hashing algorithm when new members
are added or removed.
* Send a few EndHeartbeat when ending send of Heartbeat messages.
* Due to the shutdown issues the TestConductorTransport is by
default not active, but it's easy to activate it and exception
will be thrown if trying to use the featues that require it, i.e
blackhole, passThrow and throttle
* Documented
