Everything one Application replica needs to join its deployment; the Java builder's setters, as properties.
More...
|
| const string | DefaultEgressChannel = "aeron:udp?endpoint=localhost:0" |
| | An ephemeral port: one cluster session, and no port of its own to allocate.
|
| |
|
| int | SourceId [get] |
| | This application's sourceId, the one its <applications> row declares — required, because spec §5 is one id space, and never −1, which ingress refuses.
|
| |
| IReadOnlyList< string > | AddressedGroups = Array.Empty<string>() [get] |
| | The names of the groups this application publishes to besides its own; one the log's group table lacks fences it with ClusterError.GroupUnknown.
|
| |
| required int | ClientId [get] |
| | This replica's client id, its counters' label — required, and unique among the clients on its node: two sharing one share their counters' series.
|
| |
| int | MemberId [get] |
| | The node this replica runs on: whose queue it follows, and whose leadership opens its gate.
|
| |
| bool | OffCluster [get] |
| | Whether this node is a gateway host, running no member, rather than a member: the gate then opens whoever leads, and ingress is UDP alone. Run one instance, as nothing elects between two. Not with an IpcConnectTimeoutMs, which has no IPC ingress to time.
|
| |
| string | EgressChannel = DefaultEgressChannel [get] |
| | This client's own egress endpoint; two media drivers on one host cannot both bind a port. An ephemeral port on localhost, DefaultEgressChannel, by default.
|
| |
| string | IngressEndpoints = PortLayout.IngressEndpoints() [get] |
| | The members to reach when this node is not leading, PortLayout.IngressEndpoints() by default.
|
| |
| required IApplicationListener | Listener [get] |
| |
| SnapshotOptions | Snapshots [get] |
| | Takes this replica into snapshot rounds; without it it takes part in none, whatever its topology row says, and recovers from clusterSeqNo 1.
|
| |
| Action | OnNotLeader [get] |
| | Called whenever the gate closes, before IApplicationListener.OnLeadershipChanged(false): the place for OutstandingWork.OnNotLeader. Optional; without it the listener does so.
|
| |
| Tuning | Tuning = new Tuning() [get] |
| | The stall and capacity policy.
|
| |
| IIdleStrategy | IdleStrategy [get] |
| | What a call that blocks waits on between attempts: a publish that spins through an election, up to 10 s, and the cluster connect in Start. A yielding spin and Aeron's own connect wait by default. One that also does the caller's periodic work, such as writing a heartbeat, keeps it going while the call blocks.
|
| |
| long? | IpcConnectTimeoutMs [get] |
| | How long to wait for IPC ingress to the local member while it leads before using UDP, Application.DefaultIpcConnectTimeoutMs by default. Not with OffCluster.
|
| |
Everything one Application replica needs to join its deployment; the Java builder's setters, as properties.
◆ DefaultEgressChannel
| const string Org.Limitless.Seqeron.App.ApplicationOptions.DefaultEgressChannel = "aeron:udp?endpoint=localhost:0" |
|
static |
An ephemeral port: one cluster session, and no port of its own to allocate.
◆ AddressedGroups
| IReadOnlyList<string> Org.Limitless.Seqeron.App.ApplicationOptions.AddressedGroups = Array.Empty<string>() |
|
get |
The names of the groups this application publishes to besides its own; one the log's group table lacks fences it with ClusterError.GroupUnknown.
◆ ClientId
| required int Org.Limitless.Seqeron.App.ApplicationOptions.ClientId |
|
get |
This replica's client id, its counters' label — required, and unique among the clients on its node: two sharing one share their counters' series.
◆ EgressChannel
This client's own egress endpoint; two media drivers on one host cannot both bind a port. An ephemeral port on localhost, DefaultEgressChannel, by default.
◆ IdleStrategy
| IIdleStrategy Org.Limitless.Seqeron.App.ApplicationOptions.IdleStrategy |
|
get |
What a call that blocks waits on between attempts: a publish that spins through an election, up to 10 s, and the cluster connect in Start. A yielding spin and Aeron's own connect wait by default. One that also does the caller's periodic work, such as writing a heartbeat, keeps it going while the call blocks.
◆ IngressEndpoints
| string Org.Limitless.Seqeron.App.ApplicationOptions.IngressEndpoints = PortLayout.IngressEndpoints() |
|
get |
◆ IpcConnectTimeoutMs
| long? Org.Limitless.Seqeron.App.ApplicationOptions.IpcConnectTimeoutMs |
|
get |
How long to wait for IPC ingress to the local member while it leads before using UDP, Application.DefaultIpcConnectTimeoutMs by default. Not with OffCluster.
◆ Listener
◆ MemberId
| int Org.Limitless.Seqeron.App.ApplicationOptions.MemberId |
|
get |
The node this replica runs on: whose queue it follows, and whose leadership opens its gate.
◆ OffCluster
| bool Org.Limitless.Seqeron.App.ApplicationOptions.OffCluster |
|
get |
Whether this node is a gateway host, running no member, rather than a member: the gate then opens whoever leads, and ingress is UDP alone. Run one instance, as nothing elects between two. Not with an IpcConnectTimeoutMs, which has no IPC ingress to time.
◆ OnNotLeader
| Action Org.Limitless.Seqeron.App.ApplicationOptions.OnNotLeader |
|
get |
Called whenever the gate closes, before IApplicationListener.OnLeadershipChanged(false): the place for OutstandingWork.OnNotLeader. Optional; without it the listener does so.
◆ Snapshots
Takes this replica into snapshot rounds; without it it takes part in none, whatever its topology row says, and recovers from clusterSeqNo 1.
◆ SourceId
| int Org.Limitless.Seqeron.App.ApplicationOptions.SourceId |
|
get |
This application's sourceId, the one its <applications> row declares — required, because spec §5 is one id space, and never −1, which ingress refuses.
◆ Tuning
| Tuning Org.Limitless.Seqeron.App.ApplicationOptions.Tuning = new Tuning() |
|
get |
The stall and capacity policy.
The documentation for this class was generated from the following file: