CORDA-1494: Rename p2pMessagingRetry configuration to flowTimeout configuration to reflect the change in how retries are implemented – it is now on flow rather than message level. (#3339)

This commit is contained in:
Andrius Dagys 2018-06-11 15:50:06 +01:00 committed by GitHub
parent c66228adf4
commit 85fe0e4cb1
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
9 changed files with 29 additions and 28 deletions

View File

@ -94,12 +94,14 @@ absolute path to the node's base directory.
here must be externally accessible when running nodes across a cluster of machines. If the provided host is unreachable,
the node will try to auto-discover its public one.
:p2pMessagingRetry: Only used for notarisation requests. When the response doesn't arrive in time, the message is
resent to a different notary-replica round-robin in case of clustered notaries.
:flowTimeout: When a flow implementing the ``TimedFlow`` interface does not complete in time, it is restarted from the
initial checkpoint. Currently only used for notarisation requests: if a notary replica dies while processing a notarisation request,
the client flow eventually times out and gets restarted. On restart the request is resent to a different notary replica
in a round-robin fashion (assuming the notary is clustered).
:messageRedeliveryDelay: The initial retry delay, e.g. `30 seconds`.
:maxRetryCount: How many retries to attempt.
:backoffBase: The base of the exponential backoff, `t_{wait} = messageRedeliveryDelay * backoffBase^{retryCount}`.
:timeout: The initial flow timeout period, e.g. `30 seconds`.
:maxRestartCount: Maximum number of times the flow will restart before resulting in an error.
:backoffBase: The base of the exponential backoff, `t_{wait} = timeout * backoffBase^{retryCount}`.
:rpcAddress: The address of the RPC system on which RPC requests can be made to the node. If not provided then the node will run without RPC. This is now deprecated in favour of the ``rpcSettings`` block.

View File

@ -23,7 +23,7 @@ import net.corda.core.utilities.seconds
import net.corda.node.internal.StartedNode
import net.corda.node.services.config.NodeConfiguration
import net.corda.node.services.config.NotaryConfig
import net.corda.node.services.config.P2PMessagingRetryConfiguration
import net.corda.node.services.config.FlowTimeoutConfiguration
import net.corda.nodeapi.internal.DevIdentityGenerator
import net.corda.nodeapi.internal.network.NetworkParametersCopier
import net.corda.testing.common.internal.testNetworkParameters
@ -101,8 +101,8 @@ class TimedFlowTests {
InternalMockNodeParameters(
legalName = CordaX500Name("Alice", "AliceCorp", "GB"),
configOverrides = { conf: NodeConfiguration ->
val retryConfig = P2PMessagingRetryConfiguration(1.seconds, 3, 1.0)
doReturn(retryConfig).whenever(conf).p2pMessagingRetry
val retryConfig = FlowTimeoutConfiguration(1.seconds, 3, 1.0)
doReturn(retryConfig).whenever(conf).flowTimeout
}
)
)

View File

@ -39,7 +39,7 @@ interface NodeConfiguration : NodeSSLConfiguration {
val networkServices: NetworkServicesConfig?
val certificateChainCheckPolicies: List<CertChainPolicyConfig>
val verifierType: VerifierType
val p2pMessagingRetry: P2PMessagingRetryConfiguration
val flowTimeout: FlowTimeoutConfiguration
val notary: NotaryConfig?
val additionalNodeInfoPollingFrequencyMsec: Long
val p2pAddress: NetworkHostAndPort
@ -139,12 +139,11 @@ data class NetworkServicesConfig(
/**
* Currently only used for notarisation requests.
*
* When the response doesn't arrive in time, the message is resent to a different notary-replica round-robin
* in case of clustered notaries.
* Specifies the configuration for timing out and restarting a [TimedFlow].
*/
data class P2PMessagingRetryConfiguration(
val messageRedeliveryDelay: Duration,
val maxRetryCount: Int,
data class FlowTimeoutConfiguration(
val timeout: Duration,
val maxRestartCount: Int,
val backoffBase: Double
)
@ -167,7 +166,7 @@ data class NodeConfigurationImpl(
override val rpcUsers: List<User>,
override val security: SecurityConfiguration? = null,
override val verifierType: VerifierType,
override val p2pMessagingRetry: P2PMessagingRetryConfiguration,
override val flowTimeout: FlowTimeoutConfiguration,
override val p2pAddress: NetworkHostAndPort,
private val rpcAddress: NetworkHostAndPort? = null,
private val rpcSettings: NodeRpcSettings,

View File

@ -609,10 +609,10 @@ class SingleThreadedStateMachineManager(
/** Schedules a [FlowTimeoutException] to be fired in order to restart the flow. */
private fun scheduleTimeoutException(flow: Flow, retryCount: Int): ScheduledFuture<*> {
return with(serviceHub.configuration.p2pMessagingRetry) {
val timeoutDelaySeconds = messageRedeliveryDelay.seconds * Math.pow(backoffBase, retryCount.toDouble()).toLong()
return with(serviceHub.configuration.flowTimeout) {
val timeoutDelaySeconds = timeout.seconds * Math.pow(backoffBase, retryCount.toDouble()).toLong()
timeoutScheduler.schedule({
val event = Event.Error(FlowTimeoutException(maxRetryCount))
val event = Event.Error(FlowTimeoutException(maxRestartCount))
flow.fiber.scheduleEvent(event)
}, timeoutDelaySeconds, TimeUnit.SECONDS)
}

View File

@ -20,8 +20,8 @@ rpcSettings = {
useSsl = false
standAloneBroker = false
}
p2pMessagingRetry {
messageRedeliveryDelay = 30 seconds
maxRetryCount = 3
flowTimeout {
timeout = 30 seconds
maxRestartCount = 3
backoffBase = 2.0
}

View File

@ -229,7 +229,7 @@ class NodeConfigurationImplTest {
verifierType = VerifierType.InMemory,
p2pAddress = NetworkHostAndPort("localhost", 0),
messagingServerAddress = null,
p2pMessagingRetry = P2PMessagingRetryConfiguration(5.seconds, 3, 1.0),
flowTimeout = FlowTimeoutConfiguration(5.seconds, 3, 1.0),
notary = null,
devMode = true,
noLocalShell = false,

View File

@ -8,7 +8,7 @@ import net.corda.core.utilities.seconds
import net.corda.node.internal.configureDatabase
import net.corda.node.services.config.CertChainPolicyConfig
import net.corda.node.services.config.NodeConfiguration
import net.corda.node.services.config.P2PMessagingRetryConfiguration
import net.corda.node.services.config.FlowTimeoutConfiguration
import net.corda.node.services.config.configureWithDevSSLCertificate
import net.corda.node.services.network.NetworkMapCacheImpl
import net.corda.node.services.network.PersistentNetworkMapCache
@ -73,7 +73,7 @@ class ArtemisMessagingTest {
doReturn(NetworkHostAndPort("0.0.0.0", serverPort)).whenever(it).p2pAddress
doReturn(null).whenever(it).jmxMonitoringHttpPort
doReturn(emptyList<CertChainPolicyConfig>()).whenever(it).certificateChainCheckPolicies
doReturn(P2PMessagingRetryConfiguration(5.seconds, 3, backoffBase = 1.0)).whenever(it).p2pMessagingRetry
doReturn(FlowTimeoutConfiguration(5.seconds, 3, backoffBase = 1.0)).whenever(it).flowTimeout
}
LogHelper.setLevel(PersistentUniquenessProvider::class)
database = configureDatabase(makeTestDataSourceProperties(), DatabaseConfig(), { null }, { null })

View File

@ -24,8 +24,8 @@ rpcSettings = {
useSsl = false
standAloneBroker = false
}
p2pMessagingRetry {
messageRedeliveryDelay = 30 seconds
maxRetryCount = 3
flowTimeout {
timeout = 30 seconds
maxRestartCount = 3
backoffBase = 2.0
}

View File

@ -483,7 +483,7 @@ private fun mockNodeConfiguration(): NodeConfiguration {
doReturn(null).whenever(it).networkServices
doReturn(VerifierType.InMemory).whenever(it).verifierType
// Set to be long enough so retries don't trigger unless we override it
doReturn(P2PMessagingRetryConfiguration(1.hours, 3, backoffBase = 2.0)).whenever(it).p2pMessagingRetry
doReturn(FlowTimeoutConfiguration(1.hours, 3, backoffBase = 1.0)).whenever(it).flowTimeout
doReturn(5.seconds.toMillis()).whenever(it).additionalNodeInfoPollingFrequencyMsec
doReturn(null).whenever(it).devModeOptions
}