raboof/pekko
1
0
Fork 0
Code Issues Pull requests Projects Releases 6 Packages Wiki Activity Actions 11
pekko/akka-cluster/src/test/scala/akka/cluster/ReachabilitySpec.scala

262 lines
9.7 KiB
Scala
Raw Normal View History

update headers to regular comment (#25807)
2018-10-29 17:19:37 +08:00
/*
Update headers from 2018 to 2019 once for all. (#26165) * Add CopyrightHeader support for sbt-boilerplate plugin. * Add CopyrightHeader support for `*.proto` files. * Add regex match for both `–` and `-` for CopyrightHeader. * Add CopyrightHeader support for sbt build files. * Update copyright from 2018 to 2019.
2019-01-02 18:55:26 +08:00
* Copyright (C) 2009-2019 Lightbend Inc. <https://www.lightbend.com>
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
*/
package akka.cluster
import org.scalatest.WordSpec
=pro #3759 Changed to using non-deprecated ScalaTest Matchers
2013-12-17 14:25:56 +01:00
import org.scalatest.Matchers
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
import akka.actor.Address
=pro #3759 Changed to using non-deprecated ScalaTest Matchers
2013-12-17 14:25:56 +01:00
class ReachabilitySpec extends WordSpec with Matchers {
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
format source with scalafmt
2019-03-11 10:38:24 +01:00
import Reachability.{ Reachable, Record, Terminated, Unreachable }
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
more classic remoting leftovers * fix several tests that were using akka.remote.netty.tcp config * replace akka.tcp with akka in more places * some doc cross references to remoting.md
2019-05-15 18:01:34 +02:00
val nodeA = UniqueAddress(Address("akka", "sys", "a", 2552), 1L)
val nodeB = UniqueAddress(Address("akka", "sys", "b", 2552), 2L)
val nodeC = UniqueAddress(Address("akka", "sys", "c", 2552), 3L)
val nodeD = UniqueAddress(Address("akka", "sys", "d", 2552), 4L)
val nodeE = UniqueAddress(Address("akka", "sys", "e", 2552), 5L)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
"Reachability table" must {
"be reachable when empty" in {
val r = Reachability.empty
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.isReachable(nodeA) should ===(true)
r.allUnreachable should ===(Set.empty)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"be unreachable when one observed unreachable" in {
val r = Reachability.empty.unreachable(nodeB, nodeA)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.isReachable(nodeA) should ===(false)
r.allUnreachable should ===(Set(nodeA))
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"not be reachable when terminated" in {
val r = Reachability.empty.terminated(nodeB, nodeA)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.isReachable(nodeA) should ===(false)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
// allUnreachable doesn't include terminated
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.allUnreachable should ===(Set.empty)
r.allUnreachableOrTerminated should ===(Set(nodeA))
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"not change terminated entry" in {
val r = Reachability.empty.terminated(nodeB, nodeA)
format source with scalafmt
2019-03-11 10:38:24 +01:00
(r.reachable(nodeB, nodeA) should be).theSameInstanceAs(r)
(r.unreachable(nodeB, nodeA) should be).theSameInstanceAs(r)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"not change when same status" in {
val r = Reachability.empty.unreachable(nodeB, nodeA)
format source with scalafmt
2019-03-11 10:38:24 +01:00
(r.unreachable(nodeB, nodeA) should be).theSameInstanceAs(r)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"be unreachable when some observed unreachable and others reachable" in {
val r = Reachability.empty.unreachable(nodeB, nodeA).unreachable(nodeC, nodeA).reachable(nodeD, nodeA)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.isReachable(nodeA) should ===(false)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"be reachable when all observed reachable again" in {
format source with scalafmt
2019-03-11 10:38:24 +01:00
val r = Reachability.empty
.unreachable(nodeB, nodeA)
.unreachable(nodeC, nodeA)
.reachable(nodeB, nodeA)
.reachable(nodeC, nodeA)
.unreachable(nodeB, nodeC)
.unreachable(nodeC, nodeB)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.isReachable(nodeA) should ===(true)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
=clu #13875 Exclude unreachability observations from downed * Skip observations from downed node (quarantined is marked down immediately) in convergence check * Skip observations from downed node when picking "reachable" targets for gossip. * This also means that we must accept gossip with own node marked as unreachable, but that should not be spread to the external membership events.
2015-01-30 14:30:16 +01:00
"exclude observations from specific (downed) nodes" in {
format source with scalafmt
2019-03-11 10:38:24 +01:00
val r = Reachability.empty
.unreachable(nodeC, nodeA)
.reachable(nodeC, nodeA)
.unreachable(nodeC, nodeB)
.unreachable(nodeB, nodeA)
.unreachable(nodeB, nodeC)
=clu #13875 Exclude unreachability observations from downed * Skip observations from downed node (quarantined is marked down immediately) in convergence check * Skip observations from downed node when picking "reachable" targets for gossip. * This also means that we must accept gossip with own node marked as unreachable, but that should not be spread to the external membership events.
2015-01-30 14:30:16 +01:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.isReachable(nodeA) should ===(false)
r.isReachable(nodeB) should ===(false)
r.isReachable(nodeC) should ===(false)
r.allUnreachableOrTerminated should ===(Set(nodeA, nodeB, nodeC))
r.removeObservers(Set(nodeB)).allUnreachableOrTerminated should ===(Set(nodeB))
=clu #13875 Exclude unreachability observations from downed * Skip observations from downed node (quarantined is marked down immediately) in convergence check * Skip observations from downed node when picking "reachable" targets for gossip. * This also means that we must accept gossip with own node marked as unreachable, but that should not be spread to the external membership events.
2015-01-30 14:30:16 +01:00
}
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
"be pruned when all records of an observer are Reachable" in {
format source with scalafmt
2019-03-11 10:38:24 +01:00
val r = Reachability.empty
.unreachable(nodeB, nodeA)
.unreachable(nodeB, nodeC)
.unreachable(nodeD, nodeC)
.reachable(nodeB, nodeA)
.reachable(nodeB, nodeC)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.isReachable(nodeA) should ===(true)
r.isReachable(nodeC) should ===(false)
r.records should ===(Vector(Record(nodeD, nodeC, Unreachable, 1L)))
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
val r2 = r.unreachable(nodeB, nodeD).unreachable(nodeB, nodeE)
format source with scalafmt
2019-03-11 10:38:24 +01:00
r2.records.toSet should ===(
format source with scalafmt, #26511
2019-03-13 10:56:20 +01:00
Set(
Record(nodeD, nodeC, Unreachable, 1L),
Record(nodeB, nodeD, Unreachable, 5L),
Record(nodeB, nodeE, Unreachable, 6L)))
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"have correct aggregated status" in {
format source with scalafmt, #26511
2019-03-13 10:56:20 +01:00
val records = Vector(
Reachability.Record(nodeA, nodeB, Reachable, 2),
Reachability.Record(nodeC, nodeB, Unreachable, 2),
Reachability.Record(nodeA, nodeD, Unreachable, 3),
Reachability.Record(nodeD, nodeB, Terminated, 4))
replace unicode arrows * ⇒, →, ← * because we don't want to show them in documentation snippets and then it's complicated to avoid that when snippets are located in src/test/scala in individual modules * dont replace object `→` in FSM.scala and PersistentFSM.scala
2019-02-09 15:25:39 +01:00
val versions = Map(nodeA -> 3L, nodeC -> 3L, nodeD -> 4L)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
val r = Reachability(records, versions)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.status(nodeA) should ===(Reachable)
r.status(nodeB) should ===(Terminated)
r.status(nodeD) should ===(Unreachable)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"have correct status for a mix of nodes" in {
format source with scalafmt
2019-03-11 10:38:24 +01:00
val r = Reachability.empty
.unreachable(nodeB, nodeA)
.unreachable(nodeC, nodeA)
.unreachable(nodeD, nodeA)
.unreachable(nodeC, nodeB)
.reachable(nodeC, nodeB)
.unreachable(nodeD, nodeB)
.unreachable(nodeD, nodeC)
.reachable(nodeD, nodeC)
.reachable(nodeE, nodeD)
.unreachable(nodeA, nodeE)
.terminated(nodeB, nodeE)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.status(nodeB, nodeA) should ===(Unreachable)
r.status(nodeC, nodeA) should ===(Unreachable)
r.status(nodeD, nodeA) should ===(Unreachable)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.status(nodeC, nodeB) should ===(Reachable)
r.status(nodeD, nodeB) should ===(Unreachable)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.status(nodeA, nodeE) should ===(Unreachable)
r.status(nodeB, nodeE) should ===(Terminated)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.isReachable(nodeA) should ===(false)
r.isReachable(nodeB) should ===(false)
r.isReachable(nodeC) should ===(true)
r.isReachable(nodeD) should ===(true)
r.isReachable(nodeE) should ===(false)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.allUnreachable should ===(Set(nodeA, nodeB))
r.allUnreachableFrom(nodeA) should ===(Set(nodeE))
r.allUnreachableFrom(nodeB) should ===(Set(nodeA))
r.allUnreachableFrom(nodeC) should ===(Set(nodeA))
r.allUnreachableFrom(nodeD) should ===(Set(nodeA, nodeB))
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
format source with scalafmt
2019-03-11 10:38:24 +01:00
r.observersGroupedByUnreachable should ===(
Map(nodeA -> Set(nodeB, nodeC, nodeD), nodeB -> Set(nodeD), nodeE -> Set(nodeA)))
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"merge by picking latest version of each record" in {
val r1 = Reachability.empty.unreachable(nodeB, nodeA).unreachable(nodeC, nodeD)
val r2 = r1.reachable(nodeB, nodeA).unreachable(nodeD, nodeE).unreachable(nodeC, nodeA)
val merged = r1.merge(Set(nodeA, nodeB, nodeC, nodeD, nodeE), r2)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged.status(nodeB, nodeA) should ===(Reachable)
merged.status(nodeC, nodeA) should ===(Unreachable)
merged.status(nodeC, nodeD) should ===(Unreachable)
merged.status(nodeD, nodeE) should ===(Unreachable)
merged.status(nodeE, nodeA) should ===(Reachable)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged.isReachable(nodeA) should ===(false)
merged.isReachable(nodeD) should ===(false)
merged.isReachable(nodeE) should ===(false)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
val merged2 = r2.merge(Set(nodeA, nodeB, nodeC, nodeD, nodeE), r1)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged2.records.toSet should ===(merged.records.toSet)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
=clu #3603 Handle removed member in Gossip and Reachability merge * It was a regression introduced in dc9fe4f * Two problems: 1) Gossip merge could pop back removed member (was previously covered by the filter of unreachable) 2) Reachability merge didn't handle all cases for removed member, i.e. when node not in allowed set
2013-09-12 17:19:53 +02:00
"merge by taking allowed set into account" in {
val r1 = Reachability.empty.unreachable(nodeB, nodeA).unreachable(nodeC, nodeD)
val r2 = r1.reachable(nodeB, nodeA).unreachable(nodeD, nodeE).unreachable(nodeC, nodeA)
// nodeD not in allowed set
val allowed = Set(nodeA, nodeB, nodeC, nodeE)
val merged = r1.merge(allowed, r2)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged.status(nodeB, nodeA) should ===(Reachable)
merged.status(nodeC, nodeA) should ===(Unreachable)
merged.status(nodeC, nodeD) should ===(Reachable)
merged.status(nodeD, nodeE) should ===(Reachable)
merged.status(nodeE, nodeA) should ===(Reachable)
=clu #3603 Handle removed member in Gossip and Reachability merge * It was a regression introduced in dc9fe4f * Two problems: 1) Gossip merge could pop back removed member (was previously covered by the filter of unreachable) 2) Reachability merge didn't handle all cases for removed member, i.e. when node not in allowed set
2013-09-12 17:19:53 +02:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged.isReachable(nodeA) should ===(false)
merged.isReachable(nodeD) should ===(true)
merged.isReachable(nodeE) should ===(true)
=clu #3603 Handle removed member in Gossip and Reachability merge * It was a regression introduced in dc9fe4f * Two problems: 1) Gossip merge could pop back removed member (was previously covered by the filter of unreachable) 2) Reachability merge didn't handle all cases for removed member, i.e. when node not in allowed set
2013-09-12 17:19:53 +02:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged.versions.keySet should ===(Set(nodeB, nodeC))
=clu #3603 Handle removed member in Gossip and Reachability merge * It was a regression introduced in dc9fe4f * Two problems: 1) Gossip merge could pop back removed member (was previously covered by the filter of unreachable) 2) Reachability merge didn't handle all cases for removed member, i.e. when node not in allowed set
2013-09-12 17:19:53 +02:00
val merged2 = r2.merge(allowed, r1)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged2.records.toSet should ===(merged.records.toSet)
merged2.versions should ===(merged.versions)
=clu #3603 Handle removed member in Gossip and Reachability merge * It was a regression introduced in dc9fe4f * Two problems: 1) Gossip merge could pop back removed member (was previously covered by the filter of unreachable) 2) Reachability merge didn't handle all cases for removed member, i.e. when node not in allowed set
2013-09-12 17:19:53 +02:00
}
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
"merge correctly after pruning" in {
val r1 = Reachability.empty.unreachable(nodeB, nodeA).unreachable(nodeC, nodeD)
val r2 = r1.unreachable(nodeA, nodeE)
val r3 = r1.reachable(nodeB, nodeA) // nodeB pruned
val merged = r2.merge(Set(nodeA, nodeB, nodeC, nodeD, nodeE), r3)
format source with scalafmt
2019-03-11 10:38:24 +01:00
merged.records.toSet should ===(Set(Record(nodeA, nodeE, Unreachable, 1), Record(nodeC, nodeD, Unreachable, 1)))
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
val merged3 = r3.merge(Set(nodeA, nodeB, nodeC, nodeD, nodeE), r2)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged3.records.toSet should ===(merged.records.toSet)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"merge versions correctly" in {
replace unicode arrows * ⇒, →, ← * because we don't want to show them in documentation snippets and then it's complicated to avoid that when snippets are located in src/test/scala in individual modules * dont replace object `→` in FSM.scala and PersistentFSM.scala
2019-02-09 15:25:39 +01:00
val r1 = Reachability(Vector.empty, Map(nodeA -> 3L, nodeB -> 5L, nodeC -> 7L))
val r2 = Reachability(Vector.empty, Map(nodeA -> 6L, nodeB -> 2L, nodeD -> 1L))
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
val merged = r1.merge(Set(nodeA, nodeB, nodeC, nodeD, nodeE), r2)
replace unicode arrows * ⇒, →, ← * because we don't want to show them in documentation snippets and then it's complicated to avoid that when snippets are located in src/test/scala in individual modules * dont replace object `→` in FSM.scala and PersistentFSM.scala
2019-02-09 15:25:39 +01:00
val expected = Map(nodeA -> 6L, nodeB -> 5L, nodeC -> 7L, nodeD -> 1L)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged.versions should ===(expected)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
val merged2 = r2.merge(Set(nodeA, nodeB, nodeC, nodeD, nodeE), r1)
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
merged2.versions should ===(expected)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
"remove node" in {
format source with scalafmt
2019-03-11 10:38:24 +01:00
val r = Reachability.empty
.unreachable(nodeB, nodeA)
.unreachable(nodeC, nodeD)
.unreachable(nodeB, nodeC)
.unreachable(nodeB, nodeE)
.remove(Set(nodeA, nodeB))
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
=tes Use ConversionCheckedTripleEquals
2015-01-16 11:09:59 +01:00
r.status(nodeB, nodeA) should ===(Reachable)
r.status(nodeC, nodeD) should ===(Unreachable)
r.status(nodeB, nodeC) should ===(Reachable)
r.status(nodeB, nodeE) should ===(Reachable)
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
Reachability.remove didn't always remove all, #22012 * the versions table in Reachability was not cleared if the records for removed node had been pruned, i.e. all reachable again
2016-12-16 12:25:37 +01:00
"remove correctly after pruning" in {
format source with scalafmt
2019-03-11 10:38:24 +01:00
val r = Reachability.empty
.unreachable(nodeB, nodeA)
.unreachable(nodeB, nodeC)
.unreachable(nodeD, nodeC)
.reachable(nodeB, nodeA)
.reachable(nodeB, nodeC)
Reachability.remove didn't always remove all, #22012 * the versions table in Reachability was not cleared if the records for removed node had been pruned, i.e. all reachable again
2016-12-16 12:25:37 +01:00
r.records should ===(Vector(Record(nodeD, nodeC, Unreachable, 1L)))
val r2 = r.remove(List(nodeB))
r2.allObservers should ===(Set(nodeD))
r2.versions.keySet should ===(Set(nodeD))
}
Reachability.allObservers wrong after filterRecords, #25950
2018-11-22 13:16:35 +01:00
"be able to filter records" in {
format source with scalafmt
2019-03-11 10:38:24 +01:00
val r = Reachability.empty.unreachable(nodeC, nodeB).unreachable(nodeB, nodeA).unreachable(nodeB, nodeC)
Reachability.allObservers wrong after filterRecords, #25950
2018-11-22 13:16:35 +01:00
replace unicode arrows * ⇒, →, ← * because we don't want to show them in documentation snippets and then it's complicated to avoid that when snippets are located in src/test/scala in individual modules * dont replace object `→` in FSM.scala and PersistentFSM.scala
2019-02-09 15:25:39 +01:00
val filtered1 = r.filterRecords(record => record.observer != nodeC)
Reachability.allObservers wrong after filterRecords, #25950
2018-11-22 13:16:35 +01:00
filtered1.isReachable(nodeB) should ===(true)
filtered1.isReachable(nodeA) should ===(false)
filtered1.allObservers should ===(Set(nodeB))
}
!clu #2307 Allow transition from unreachable to reachable * Replace unreachable Set with Reachability table * Unreachable members stay in member Set * Downing a live member was moved it to the unreachable Set, and then removed from there by the leader. That will not work when flipping back to reachable, so a Down member must be detected as unreachable before beeing removed. Similar to Exiting. Member shuts down itself if it sees itself as Down. * Flip back to reachable when failure detector monitors it as available again * ReachableMember event * Can't ignore gossip from aggregated unreachable (see SurviveNetworkInstabilitySpec) * Make use of ReachableMember event in cluster router * End heartbeat when acknowledged, EndHeartbeatAck * Remove nr-of-end-heartbeats from conf * Full reachability info in JMX cluster status * Don't use interval after unreachable for AccrualFailureDetector history * Add QuarantinedEvent to remoting, used for Reachability.Terminated * Prune reachability table when all reachable * Update documentation * Performance testing and optimizations
2013-08-27 15:14:53 +02:00
}
}
Reference in a new issue Copy permalink