mirror of
https://github.com/whyour/qinglong.git
synced 2026-09-22 10:32:40 +08:00
feat(ql3): establish 3.0 incubation baseline
This commit is contained in:
@@ -0,0 +1,261 @@
|
||||
'use strict';
|
||||
|
||||
const assert = require('node:assert/strict');
|
||||
const { test } = require('node:test');
|
||||
|
||||
const {
|
||||
ClusterRunLostRetryCoordinator,
|
||||
ClusterRunLostRetryUnavailableError,
|
||||
buildClusterRunLostRetryTransition,
|
||||
normalizeClusterRunLostRetryPageCommand,
|
||||
normalizeClusterRunLostRetryPageResult,
|
||||
} = require('../dist/run/clusterRunLostRetry');
|
||||
|
||||
function lostRun(status = 'lost') {
|
||||
return {
|
||||
id: 'run-1',
|
||||
projectId: 'project-1',
|
||||
taskId: 'task-1',
|
||||
taskRevision: 'revision-1',
|
||||
triggerType: 'manual',
|
||||
executionOrigin: 'manual',
|
||||
executionOwner: 'runtime',
|
||||
status,
|
||||
version: 4,
|
||||
eventSequence: 4,
|
||||
priority: 0,
|
||||
createdAtMs: 100,
|
||||
queuedAtMs: 110,
|
||||
startedAtMs: 150,
|
||||
errorCode: 'CLUSTER_RECOVERY_EXECUTION_NOT_RUNNING',
|
||||
errorSummary: 'lost',
|
||||
};
|
||||
}
|
||||
|
||||
function lostAttempt(attempt = 1) {
|
||||
return {
|
||||
id: `attempt-${attempt}`,
|
||||
runId: 'run-1',
|
||||
attempt,
|
||||
status: 'lost',
|
||||
executorType: 'remote_worker',
|
||||
callbackSequence: 0,
|
||||
createdAtMs: 120,
|
||||
startedAtMs: 150,
|
||||
finishedAtMs: 200,
|
||||
errorCode: 'CLUSTER_RECOVERY_EXECUTION_NOT_RUNNING',
|
||||
errorSummary: 'lost',
|
||||
};
|
||||
}
|
||||
|
||||
function policy(overrides = {}) {
|
||||
return {
|
||||
runId: 'run-1',
|
||||
maxAttempts: 3,
|
||||
retryOnLost: true,
|
||||
safety: 'idempotent',
|
||||
backoffBaseMs: 1_000,
|
||||
backoffMaxMs: 8_000,
|
||||
version: 0,
|
||||
createdAtMs: 100,
|
||||
updatedAtMs: 100,
|
||||
...overrides,
|
||||
};
|
||||
}
|
||||
|
||||
test('schedules a safe lost Run from its admitted immutable policy', () => {
|
||||
const result = buildClusterRunLostRetryTransition({
|
||||
run: lostRun(),
|
||||
attempt: lostAttempt(),
|
||||
policy: policy(),
|
||||
observedAtMs: 300,
|
||||
runEventId: 'event-1',
|
||||
});
|
||||
|
||||
assert.equal(result.disposition, 'scheduled');
|
||||
assert.equal(result.runTransitions.length, 1);
|
||||
assert.equal(result.runTransitions[0].status, 'retry_wait');
|
||||
assert.equal(result.runTransitions[0].version, 5);
|
||||
assert.equal(result.policy.nextAttemptAtMs, 1_200);
|
||||
assert.equal(result.policy.version, 1);
|
||||
assert.deepEqual(
|
||||
result.events.map((event) => [event.type, event.sequence]),
|
||||
[['run.retry_wait', 5]],
|
||||
);
|
||||
});
|
||||
|
||||
test('requeues a due retry_wait Run with one fresh unleased Attempt', () => {
|
||||
const result = buildClusterRunLostRetryTransition({
|
||||
run: lostRun('retry_wait'),
|
||||
attempt: lostAttempt(),
|
||||
policy: policy({ nextAttemptAtMs: 250 }),
|
||||
observedAtMs: 300,
|
||||
runEventId: 'event-queued',
|
||||
attemptId: 'attempt-2',
|
||||
attemptEventId: 'event-claimed',
|
||||
});
|
||||
|
||||
assert.equal(result.disposition, 'requeued');
|
||||
assert.deepEqual(
|
||||
result.runTransitions.map((run) => [run.status, run.version]),
|
||||
[
|
||||
['queued', 5],
|
||||
['queued', 6],
|
||||
],
|
||||
);
|
||||
assert.deepEqual(result.attempt, {
|
||||
id: 'attempt-2',
|
||||
runId: 'run-1',
|
||||
attempt: 2,
|
||||
status: 'claimed',
|
||||
executorType: 'remote_worker',
|
||||
callbackSequence: 0,
|
||||
createdAtMs: 300,
|
||||
});
|
||||
assert.equal(
|
||||
Object.prototype.hasOwnProperty.call(result.policy, 'nextAttemptAtMs'),
|
||||
false,
|
||||
);
|
||||
assert.deepEqual(
|
||||
result.events.map((event) => [event.type, event.sequence]),
|
||||
[
|
||||
['run.queued', 5],
|
||||
['attempt.claimed', 6],
|
||||
],
|
||||
);
|
||||
});
|
||||
|
||||
test('terminalizes disabled, unsafe and exhausted policies without retrying', () => {
|
||||
const cases = [
|
||||
[null, lostAttempt(), 'failed_disabled', 'RUN_LOST_RETRY_DISABLED'],
|
||||
[
|
||||
policy({ safety: 'unknown' }),
|
||||
lostAttempt(),
|
||||
'failed_unsafe',
|
||||
'RUN_LOST_RETRY_UNSAFE',
|
||||
],
|
||||
[
|
||||
policy({ maxAttempts: 3 }),
|
||||
lostAttempt(3),
|
||||
'failed_exhausted',
|
||||
'RUN_LOST_RETRY_EXHAUSTED',
|
||||
],
|
||||
];
|
||||
for (const [admittedPolicy, attempt, disposition, errorCode] of cases) {
|
||||
const result = buildClusterRunLostRetryTransition({
|
||||
run: lostRun(),
|
||||
attempt,
|
||||
policy: admittedPolicy,
|
||||
observedAtMs: 300,
|
||||
runEventId: `event-${disposition}`,
|
||||
});
|
||||
assert.equal(result.disposition, disposition);
|
||||
assert.equal(result.runTransitions[0].status, 'failed');
|
||||
assert.equal(result.runTransitions[0].errorCode, errorCode);
|
||||
assert.equal(result.runTransitions[0].finishedAtMs, 300);
|
||||
}
|
||||
});
|
||||
|
||||
test('rejects cancellation, Workflow aggregates and early retry_wait', () => {
|
||||
assert.throws(() =>
|
||||
buildClusterRunLostRetryTransition({
|
||||
run: { ...lostRun(), cancelRequestedAtMs: 250, cancelReason: 'user' },
|
||||
attempt: lostAttempt(),
|
||||
policy: policy(),
|
||||
observedAtMs: 300,
|
||||
runEventId: 'event-cancelled',
|
||||
}),
|
||||
);
|
||||
assert.throws(() =>
|
||||
buildClusterRunLostRetryTransition({
|
||||
run: { ...lostRun(), triggerType: 'plugin_package_workflow' },
|
||||
attempt: lostAttempt(),
|
||||
policy: policy(),
|
||||
observedAtMs: 300,
|
||||
runEventId: 'event-workflow',
|
||||
}),
|
||||
);
|
||||
assert.throws(() =>
|
||||
buildClusterRunLostRetryTransition({
|
||||
run: lostRun('retry_wait'),
|
||||
attempt: lostAttempt(),
|
||||
policy: policy({ nextAttemptAtMs: 301 }),
|
||||
observedAtMs: 300,
|
||||
runEventId: 'event-early',
|
||||
attemptId: 'attempt-2',
|
||||
attemptEventId: 'event-2',
|
||||
}),
|
||||
);
|
||||
});
|
||||
|
||||
test('normalizes one bounded page and coalesces overlapping callers', async () => {
|
||||
assert.deepEqual(normalizeClusterRunLostRetryPageCommand({ limit: 2 }), {
|
||||
limit: 2,
|
||||
});
|
||||
assert.deepEqual(
|
||||
normalizeClusterRunLostRetryPageResult(
|
||||
{
|
||||
scanned: 2,
|
||||
scheduled: 1,
|
||||
requeued: 0,
|
||||
failed: 0,
|
||||
raced: 1,
|
||||
hasMore: false,
|
||||
},
|
||||
2,
|
||||
),
|
||||
{
|
||||
scanned: 2,
|
||||
scheduled: 1,
|
||||
requeued: 0,
|
||||
failed: 0,
|
||||
raced: 1,
|
||||
hasMore: false,
|
||||
},
|
||||
);
|
||||
|
||||
let release;
|
||||
const coordinator = new ClusterRunLostRetryCoordinator(
|
||||
{
|
||||
async reconcilePage(command) {
|
||||
await new Promise((resolve) => {
|
||||
release = resolve;
|
||||
});
|
||||
return {
|
||||
scanned: command.limit,
|
||||
scheduled: command.limit,
|
||||
requeued: 0,
|
||||
failed: 0,
|
||||
raced: 0,
|
||||
hasMore: false,
|
||||
};
|
||||
},
|
||||
},
|
||||
{ pageSize: 2 },
|
||||
);
|
||||
const first = coordinator.reconcile();
|
||||
const second = coordinator.reconcile();
|
||||
assert.equal(first, second);
|
||||
while (!release) await new Promise((resolve) => setImmediate(resolve));
|
||||
release();
|
||||
assert.equal((await first).scheduled, 2);
|
||||
});
|
||||
|
||||
test('wraps malformed repository results as unavailable', async () => {
|
||||
const coordinator = new ClusterRunLostRetryCoordinator({
|
||||
async reconcilePage() {
|
||||
return {
|
||||
scanned: 1,
|
||||
scheduled: 1,
|
||||
requeued: 1,
|
||||
failed: 0,
|
||||
raced: 0,
|
||||
hasMore: false,
|
||||
};
|
||||
},
|
||||
});
|
||||
await assert.rejects(
|
||||
coordinator.reconcile(),
|
||||
ClusterRunLostRetryUnavailableError,
|
||||
);
|
||||
});
|
||||
Reference in New Issue
Block a user