refactor(#2441): move agent field from DAG onto Node; drop dedup
Agent was a single field on Dag/DagSpec, making a DAG structurally one-agent — a multi-agent op could only ever be N separate DAGs. Move it onto Node/NodeSpec (and the NodeView wire type), drop it from Dag/DagSpec (and DagView): a DAG can now span agents. - lifecycle lease keys on the node's agent, still globally exclusive per agent across all DAGs (Inner.leases unchanged in shape). A DAG holds one lease per distinct agent it touches; settle() frees each at DAG-terminal (per-agent-subgraph early release is a follow-up, only observable with multi-agent DAGs). - transient guard keyed (dag_id, agent); cancel-revert + Rebuilt events walk TerminalDag.agents. - submit-time dedup removed (a multi-agent DAG has no single agent to key on); every submit enqueues a fresh DAG. Whether dedup needs reintroducing is tracked in a follow-up sub-issue. - templates gain a node(agent, kind, deps) helper stamping the agent onto every node; meta templates stamp "hyperhive". Templates stay single-agent in this PR — behaviour is unchanged, only the representation + wire shape. Multi-agent DAG emission (restart/restart-all/ broad stop+start as one DAG) and the SetWanted-as-a-node change are follow-ups off #2439.
This commit is contained in:
atlas
mara
parent
91607f3896
commit
2a59f2f5fc
7 changed files with 223 additions and 356 deletions
|
|
@ -3,6 +3,11 @@
|
|||
//! confined to this validation; the runtime store stays the plain
|
||||
//! `Vec<Node>` + `deps`).
|
||||
//!
|
||||
//! Every node carries its own `agent` (there is no DAG-level agent) — the
|
||||
//! `node` helper stamps the template's agent onto each. Today's templates
|
||||
//! are single-agent (every node shares one agent); a future multi-agent
|
||||
//! template would stamp different agents per subgraph.
|
||||
//!
|
||||
//! ```text
|
||||
//! rebuild(a): Prebuild(a) → StopForUpdate(a) → Swap(a) →(any) Reconcile(a)
|
||||
//! graceful-stop(a): [wanted=Offline] Signal(a) → Drain(a) → Reconcile(a)
|
||||
|
|
@ -29,35 +34,42 @@ fn after_ok(on: u32) -> Vec<Dep> {
|
|||
}]
|
||||
}
|
||||
|
||||
/// Build one node targeting `agent`. The single place templates stamp a
|
||||
/// node's agent, so a whole template is single-agent by passing the same
|
||||
/// `agent` to every call.
|
||||
fn node(agent: &str, kind: NodeKind, deps: Vec<Dep>) -> NodeSpec {
|
||||
NodeSpec {
|
||||
agent: agent.to_owned(),
|
||||
kind,
|
||||
deps,
|
||||
}
|
||||
}
|
||||
|
||||
/// The rebuild node chain. `Reconcile` deps on `Swap` with `AfterAny`:
|
||||
/// it must run even when the profile swap failed, so a previously-up
|
||||
/// agent comes back on its old config (today's recovery-start). This
|
||||
/// is the only `AfterAny` edge in v1.
|
||||
fn rebuild_nodes(relock: bool, base: u32) -> Vec<NodeSpec> {
|
||||
fn rebuild_nodes(agent: &str, relock: bool, base: u32) -> Vec<NodeSpec> {
|
||||
vec![
|
||||
NodeSpec {
|
||||
kind: NodeKind::Prebuild { relock },
|
||||
deps: if base == 0 {
|
||||
node(
|
||||
agent,
|
||||
NodeKind::Prebuild { relock },
|
||||
if base == 0 {
|
||||
Vec::new()
|
||||
} else {
|
||||
after_ok(base - 1)
|
||||
},
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::StopForUpdate,
|
||||
deps: after_ok(base),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::Swap,
|
||||
deps: after_ok(base + 1),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::Reconcile,
|
||||
deps: vec![Dep {
|
||||
),
|
||||
node(agent, NodeKind::StopForUpdate, after_ok(base)),
|
||||
node(agent, NodeKind::Swap, after_ok(base + 1)),
|
||||
node(
|
||||
agent,
|
||||
NodeKind::Reconcile,
|
||||
vec![Dep {
|
||||
on: base + 2,
|
||||
when: DepWhen::AfterAny,
|
||||
}],
|
||||
},
|
||||
),
|
||||
]
|
||||
}
|
||||
|
||||
|
|
@ -75,7 +87,6 @@ pub fn rebuild(
|
|||
) -> DagSpec {
|
||||
DagSpec {
|
||||
template: Template::Rebuild,
|
||||
agent: agent.to_owned(),
|
||||
source,
|
||||
reason,
|
||||
parent_id,
|
||||
|
|
@ -83,7 +94,7 @@ pub fn rebuild(
|
|||
inputs: Vec::new(),
|
||||
perm_payload: None,
|
||||
transient: Some(TransientKind::Rebuilding),
|
||||
nodes: rebuild_nodes(relock, 0),
|
||||
nodes: rebuild_nodes(agent, relock, 0),
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -93,7 +104,6 @@ pub fn rebuild(
|
|||
pub fn approval_deploy(agent: &str, approval_id: i64, reason: String) -> DagSpec {
|
||||
DagSpec {
|
||||
template: Template::Rebuild,
|
||||
agent: agent.to_owned(),
|
||||
source: Source::Approval,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -101,10 +111,7 @@ pub fn approval_deploy(agent: &str, approval_id: i64, reason: String) -> DagSpec
|
|||
inputs: Vec::new(),
|
||||
perm_payload: None,
|
||||
transient: Some(TransientKind::Rebuilding),
|
||||
nodes: vec![NodeSpec {
|
||||
kind: NodeKind::ApprovalDeploy,
|
||||
deps: Vec::new(),
|
||||
}],
|
||||
nodes: vec![node(agent, NodeKind::ApprovalDeploy, Vec::new())],
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -117,7 +124,6 @@ pub fn approval_deploy(agent: &str, approval_id: i64, reason: String) -> DagSpec
|
|||
pub fn graceful_stop(agent: &str, source: Source, reason: String) -> DagSpec {
|
||||
DagSpec {
|
||||
template: Template::GracefulStop,
|
||||
agent: agent.to_owned(),
|
||||
source,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -126,18 +132,9 @@ pub fn graceful_stop(agent: &str, source: Source, reason: String) -> DagSpec {
|
|||
perm_payload: None,
|
||||
transient: Some(TransientKind::Stopping),
|
||||
nodes: vec![
|
||||
NodeSpec {
|
||||
kind: NodeKind::Signal,
|
||||
deps: Vec::new(),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::Drain,
|
||||
deps: after_ok(0),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::Reconcile,
|
||||
deps: after_ok(1),
|
||||
},
|
||||
node(agent, NodeKind::Signal, Vec::new()),
|
||||
node(agent, NodeKind::Drain, after_ok(0)),
|
||||
node(agent, NodeKind::Reconcile, after_ok(1)),
|
||||
],
|
||||
}
|
||||
}
|
||||
|
|
@ -148,7 +145,6 @@ pub fn graceful_stop(agent: &str, source: Source, reason: String) -> DagSpec {
|
|||
pub fn restart(agent: &str, source: Source, reason: String) -> DagSpec {
|
||||
DagSpec {
|
||||
template: Template::Restart,
|
||||
agent: agent.to_owned(),
|
||||
source,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -157,14 +153,8 @@ pub fn restart(agent: &str, source: Source, reason: String) -> DagSpec {
|
|||
perm_payload: None,
|
||||
transient: Some(TransientKind::Restarting),
|
||||
nodes: vec![
|
||||
NodeSpec {
|
||||
kind: NodeKind::StopForUpdate,
|
||||
deps: Vec::new(),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::Reconcile,
|
||||
deps: after_ok(0),
|
||||
},
|
||||
node(agent, NodeKind::StopForUpdate, Vec::new()),
|
||||
node(agent, NodeKind::Reconcile, after_ok(0)),
|
||||
],
|
||||
}
|
||||
}
|
||||
|
|
@ -180,7 +170,6 @@ pub fn restart(agent: &str, source: Source, reason: String) -> DagSpec {
|
|||
pub fn graceful_restart(agent: &str, source: Source, reason: String) -> DagSpec {
|
||||
DagSpec {
|
||||
template: Template::GracefulRestart,
|
||||
agent: agent.to_owned(),
|
||||
source,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -189,22 +178,10 @@ pub fn graceful_restart(agent: &str, source: Source, reason: String) -> DagSpec
|
|||
perm_payload: None,
|
||||
transient: Some(TransientKind::Restarting),
|
||||
nodes: vec![
|
||||
NodeSpec {
|
||||
kind: NodeKind::Signal,
|
||||
deps: Vec::new(),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::Drain,
|
||||
deps: after_ok(0),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::StopForUpdate,
|
||||
deps: after_ok(1),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::Reconcile,
|
||||
deps: after_ok(2),
|
||||
},
|
||||
node(agent, NodeKind::Signal, Vec::new()),
|
||||
node(agent, NodeKind::Drain, after_ok(0)),
|
||||
node(agent, NodeKind::StopForUpdate, after_ok(1)),
|
||||
node(agent, NodeKind::Reconcile, after_ok(2)),
|
||||
],
|
||||
}
|
||||
}
|
||||
|
|
@ -220,7 +197,6 @@ pub fn reconcile_only(
|
|||
) -> DagSpec {
|
||||
DagSpec {
|
||||
template,
|
||||
agent: agent.to_owned(),
|
||||
source,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -228,10 +204,7 @@ pub fn reconcile_only(
|
|||
inputs: Vec::new(),
|
||||
perm_payload: None,
|
||||
transient,
|
||||
nodes: vec![NodeSpec {
|
||||
kind: NodeKind::Reconcile,
|
||||
deps: Vec::new(),
|
||||
}],
|
||||
nodes: vec![node(agent, NodeKind::Reconcile, Vec::new())],
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -242,7 +215,6 @@ pub fn reconcile_only(
|
|||
pub fn spawn(agent: &str, approval_id: i64, reason: String) -> DagSpec {
|
||||
DagSpec {
|
||||
template: Template::Spawn,
|
||||
agent: agent.to_owned(),
|
||||
source: Source::Approval,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -251,22 +223,10 @@ pub fn spawn(agent: &str, approval_id: i64, reason: String) -> DagSpec {
|
|||
perm_payload: None,
|
||||
transient: Some(TransientKind::Spawning),
|
||||
nodes: vec![
|
||||
NodeSpec {
|
||||
kind: NodeKind::Provision,
|
||||
deps: Vec::new(),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::Create,
|
||||
deps: after_ok(0),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::WriteDropin,
|
||||
deps: after_ok(1),
|
||||
},
|
||||
NodeSpec {
|
||||
kind: NodeKind::Reconcile,
|
||||
deps: after_ok(2),
|
||||
},
|
||||
node(agent, NodeKind::Provision, Vec::new()),
|
||||
node(agent, NodeKind::Create, after_ok(0)),
|
||||
node(agent, NodeKind::WriteDropin, after_ok(1)),
|
||||
node(agent, NodeKind::Reconcile, after_ok(2)),
|
||||
],
|
||||
}
|
||||
}
|
||||
|
|
@ -275,14 +235,10 @@ pub fn spawn(agent: &str, approval_id: i64, reason: String) -> DagSpec {
|
|||
/// the updated `HIVE_TOOL_GROUPS` / `HIVE_CAPABILITIES` env var takes
|
||||
/// effect in the container.
|
||||
pub fn perm_change(agent: &str, source: Source, reason: String, payload: PermPayload) -> DagSpec {
|
||||
let mut nodes = vec![NodeSpec {
|
||||
kind: NodeKind::WritePermFile,
|
||||
deps: Vec::new(),
|
||||
}];
|
||||
nodes.extend(rebuild_nodes(true, 1));
|
||||
let mut nodes = vec![node(agent, NodeKind::WritePermFile, Vec::new())];
|
||||
nodes.extend(rebuild_nodes(agent, true, 1));
|
||||
DagSpec {
|
||||
template: Template::PermChange,
|
||||
agent: agent.to_owned(),
|
||||
source,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -306,7 +262,6 @@ pub fn meta_update(
|
|||
) -> DagSpec {
|
||||
DagSpec {
|
||||
template: Template::MetaUpdate,
|
||||
agent: "hyperhive".to_owned(),
|
||||
source,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -314,19 +269,17 @@ pub fn meta_update(
|
|||
inputs,
|
||||
perm_payload: None,
|
||||
transient: None,
|
||||
nodes: vec![NodeSpec {
|
||||
kind: NodeKind::MetaLock {
|
||||
nodes: vec![node(
|
||||
"hyperhive",
|
||||
NodeKind::MetaLock {
|
||||
sweep: false,
|
||||
fanout: None,
|
||||
},
|
||||
deps: Vec::new(),
|
||||
}],
|
||||
Vec::new(),
|
||||
)],
|
||||
}
|
||||
}
|
||||
|
||||
/// Boot-time sweep parent: bump meta's hyperhive input (non-fatal),
|
||||
/// then fan out `Rebuild` children for the precomputed stale agent
|
||||
/// list (topology-sorted by the caller).
|
||||
/// Boot-time root anchor DAG: a single [`NodeKind::Noop`] node that groups
|
||||
/// this boot's `StartupSweep` + per-agent `Reconcile` child DAGs (linked via
|
||||
/// `parent_id`) into one tree so the dashboard renders the boot as one entry.
|
||||
|
|
@ -336,7 +289,6 @@ pub fn meta_update(
|
|||
pub fn boot_root(reason: String) -> DagSpec {
|
||||
DagSpec {
|
||||
template: Template::Boot,
|
||||
agent: "hyperhive".to_owned(),
|
||||
source: Source::AutoUpdate,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -344,17 +296,16 @@ pub fn boot_root(reason: String) -> DagSpec {
|
|||
inputs: Vec::new(),
|
||||
perm_payload: None,
|
||||
transient: None,
|
||||
nodes: vec![NodeSpec {
|
||||
kind: NodeKind::Noop,
|
||||
deps: Vec::new(),
|
||||
}],
|
||||
nodes: vec![node("hyperhive", NodeKind::Noop, Vec::new())],
|
||||
}
|
||||
}
|
||||
|
||||
/// Boot-time sweep parent: bump meta's hyperhive input (non-fatal),
|
||||
/// then fan out `Rebuild` children for the precomputed stale agent
|
||||
/// list (topology-sorted by the caller).
|
||||
pub fn startup_sweep(reason: String, stale_agents: Vec<String>) -> DagSpec {
|
||||
DagSpec {
|
||||
template: Template::StartupSweep,
|
||||
agent: "hyperhive".to_owned(),
|
||||
source: Source::AutoUpdate,
|
||||
reason,
|
||||
parent_id: None,
|
||||
|
|
@ -362,13 +313,14 @@ pub fn startup_sweep(reason: String, stale_agents: Vec<String>) -> DagSpec {
|
|||
inputs: Vec::new(),
|
||||
perm_payload: None,
|
||||
transient: None,
|
||||
nodes: vec![NodeSpec {
|
||||
kind: NodeKind::MetaLock {
|
||||
nodes: vec![node(
|
||||
"hyperhive",
|
||||
NodeKind::MetaLock {
|
||||
sweep: true,
|
||||
fanout: Some(stale_agents),
|
||||
},
|
||||
deps: Vec::new(),
|
||||
}],
|
||||
Vec::new(),
|
||||
)],
|
||||
}
|
||||
}
|
||||
|
||||
|
|
|
|||
Loading…
Reference in a new issue