Compare commits

..
4 Commits
Author SHA1 Message Date
gitea-actions 4a929f7e7c chore(release): v1.0.11
Build and Push Images / Build jorgecuadros-web (push) Successful in 2m1s
Build and Push Images / Build jorgecuadros-api (push) Successful in 2m44s
Cut by rmancinas via the "Cut release" workflow. Pushing the tag triggers build.yml; deploy separately with tag=1.0.11.
2026-08-04 00:55:49 +00:00
rmancinasandClaude Opus 5 66d0d071b0 feat(ops): show step progress for reimport and sync jobs
Build and Push Images / Build jorgecuadros-web (push) Successful in 1m42s
Build and Push Images / Build jorgecuadros-api (push) Successful in 2m18s
A REIMPORT takes ~110 seconds and, until now, showed only a scrolling log —
there was no way to tell "halfway" from "wedged", which mattered the day one
actually did wedge.

run_all.py emits "[paso i/N] name" before each step and the API derives
progress from the job log. Emitting the marker from the Python rather than
having the UI count STEPS itself means the step count is stated in exactly
one place; adding a step cannot desync the display. Progress is derived, not
stored, for the same reason: the log is already the record of what happened,
and a separate counter could contradict it, which is precisely the confusion
a progress display exists to remove.

While RUNNING, step i is IN PROGRESS rather than finished, so only i-1 count
as done. Counting i would show 100% while the final step was still working —
and the final step (blob_extract) is the slowest, so the bar would sit at
"100%" for the longest stretch of the job.

BACKUP and RESTORE are a single mysqldump with no steps and deliberately
render no bar; a fabricated percentage would be worse than none. The safety
backup that precedes a REIMPORT is likewise named explicitly instead of
showing 0%, which reads as stuck.

Pinned by job-progress.spec.ts, including the literal line run_all.py emits,
so a change to the Python format fails a test rather than silently blanking
the panel.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-03 17:53:10 -07:00
gitea-actions f269dc8bfa chore(release): v1.0.10
Build and Push Images / Build jorgecuadros-web (push) Successful in 2m28s
Build and Push Images / Build jorgecuadros-api (push) Successful in 2m55s
Cut by rmancinas via the "Cut release" workflow. Pushing the tag triggers build.yml; deploy separately with tag=1.0.10.
2026-08-04 00:43:02 +00:00
rmancinasandClaude Opus 5 eef9a5f4c8 fix(ops): stop the replica field parser reading the next line
Build and Push Images / Build jorgecuadros-api (push) Canceled after 51s
Build and Push Images / Build jorgecuadros-web (push) Canceled after 50s
The panel reported "Error SQL: Replicate_Ignore_Server_Ids:" against a
replica that was healthy — both threads running, zero lag.

`\s` matches newlines in JavaScript, so `^\s*NAME:\s*(.*)$` let the `\s*`
after the colon walk past an EMPTY field's line break and capture the
following line. Last_SQL_Error is blank on a healthy replica and
Replicate_Ignore_Server_Ids happens to be printed immediately after it, so
the blank error field returned the next field's name as its value. Every
empty field was affected; the visible damage was that a healthy replica
rendered as broken, which is the worst direction for a health panel to fail.

Fixed with `[^\S\n]` — horizontal whitespace only — on both sides of the
field name.

Extracted as replicaField() and pinned by replication.spec.ts against the
verbatim output of the live replica, keeping the empty Last_SQL_Error
adjacent to Replicate_Ignore_Server_Ids because that exact adjacency is what
broke. Also covers the literal "NULL" lag surviving as a distinct value from
empty, and a field name that is a suffix of another (Last_Error vs
Last_SQL_Error) not matching the wrong line.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-03 17:40:49 -07:00
11 changed files with 330 additions and 13 deletions
+1 -1
View File
@@ -1,6 +1,6 @@
{
"name": "@jorgecuadros/api",
"version": "1.0.9",
"version": "1.0.11",
"private": true,
"scripts": {
"build": "nest build",
+76
View File
@@ -0,0 +1,76 @@
import { jobProgress } from "./ops.service";
/** Shape run_all.py emits, with the shell trace lines it interleaves. */
const line = (i: number, n: number, name: string) =>
`[paso ${i}/${n}] ${name}\n+ /repo/migration/.venv/bin/python /repo/migration/${name} --env prod\n[${name}] target env: prod\n validation: OK`;
describe("jobProgress", () => {
it("returns null before any step marker appears", () => {
// The safety backup runs before run_all.py, so this is the real state for
// the first stretch of every REIMPORT.
expect(jobProgress("== Respaldo de seguridad previo ==\ntablas capturadas: 39", "RUNNING")).toBeNull();
});
it("returns null for jobs that have no steps at all", () => {
// BACKUP/RESTORE are a single mysqldump; a fabricated percentage would be
// worse than none.
expect(jobProgress("mysqldump ... done", "SUCCESS")).toBeNull();
});
it("tracks the most recent marker, not the first", () => {
const log = [line(1, 9, "transform_customers.py"), line(2, 9, "transform_properties.py")].join("\n");
const p = jobProgress(log, "RUNNING");
expect(p).toMatchObject({ step: 2, total: 9, name: "transform_properties.py" });
});
/**
* The point of the whole feature. While RUNNING, step i is IN PROGRESS, so
* only i-1 are done. Counting i as complete would show 100% while the final
* and slowest step (blob_extract) is still working.
*/
it("does not claim a running step is finished", () => {
expect(jobProgress(line(1, 9, "transform_customers.py"), "RUNNING")?.percent).toBe(0);
expect(jobProgress(line(9, 9, "blob_extract.py"), "RUNNING")?.percent).toBe(88);
});
it("reaches 100 only once the job is no longer running", () => {
expect(jobProgress(line(9, 9, "blob_extract.py"), "SUCCESS")?.percent).toBe(100);
});
/** A job that died mid-way must report where it died, not 100%. */
it("reports the failed step rather than completion", () => {
const p = jobProgress(line(5, 9, "transform_transactions.py"), "FAILED");
expect(p).toMatchObject({ step: 5, total: 9 });
expect(p!.percent).toBe(55);
});
it("handles the 8-step SYNC list as well as the 9-step REIMPORT one", () => {
expect(jobProgress(line(8, 8, "transform_bank.py"), "SUCCESS")?.percent).toBe(100);
expect(jobProgress(line(4, 8, "transform_policies.py"), "RUNNING")?.percent).toBe(37);
});
/**
* Captured verbatim from `run_all.run(..., step=8, total=9)`. This is the
* contract between the Python and this parser; if run_all.py's format
* changes, this fails rather than the panel silently showing no progress.
*/
it("parses the exact line run_all.py emits", () => {
const real =
"[paso 8/9] transform_bank.py\n+ /repo/migration/.venv/bin/python /repo/migration/transform_bank.py --env prod";
expect(jobProgress(real, "RUNNING")).toMatchObject({
step: 8,
total: 9,
name: "transform_bank.py",
percent: 77,
});
});
it("ignores a malformed marker instead of reporting NaN", () => {
expect(jobProgress("[paso 3/0] x.py", "RUNNING")).toBeNull();
});
/** The marker must be at line start so log text quoting it cannot spoof it. */
it("does not match a marker embedded mid-line", () => {
expect(jobProgress("some output mentioning [paso 4/9] fake.py", "RUNNING")).toBeNull();
});
});
+52 -1
View File
@@ -212,7 +212,9 @@ export class OpsService implements OnModuleInit {
async getJob(id: string) {
const job = await this.prisma.opsJob.findUnique({ where: { id } });
if (!job) throw new NotFoundException("Trabajo no encontrado.");
return job;
// Derived, never stored: the log is the single source of truth for how far
// a job got, so progress cannot drift out of sync with it.
return { ...job, progress: jobProgress(job.log, job.status) };
}
/**
@@ -514,3 +516,52 @@ export class OpsService implements OnModuleInit {
function shq(v: string): string {
return `'${v.replace(/'/g, `'\\''`)}'`;
}
/** Progress derived from a job's log. Null when the job reports no steps. */
export interface JobProgress {
/** 1-based index of the step currently running (or last reached). */
step: number;
total: number;
/** Script name, e.g. "transform_bank.py". */
name: string;
/** 0..100, floored. 100 only once the job is no longer RUNNING. */
percent: number;
}
/**
* Parse the "[paso i/N] name" markers migration/run_all.py emits.
*
* Progress is DERIVED from the log rather than tracked in a column: the log is
* already the record of what happened, and a separate counter could disagree
* with it — which is exactly the confusion a progress display is supposed to
* remove. run_all.py owns the step count, so adding a step cannot desync this.
*
* BACKUP and RESTORE are a single mysqldump with no steps, so they return null
* and the UI shows an indeterminate spinner. Reporting a fabricated percentage
* for them would be worse than showing none.
*/
export function jobProgress(
log: string,
status: string,
): JobProgress | null {
// Last marker wins: the log grows, and the newest line is the current step.
const matches = [...log.matchAll(/^\[paso (\d+)\/(\d+)\] (\S+)/gm)];
const last = matches[matches.length - 1];
if (!last) return null;
const step = Number(last[1]);
const total = Number(last[2]);
if (!Number.isFinite(step) || !Number.isFinite(total) || total <= 0) return null;
// While RUNNING, step i means i is IN PROGRESS, not finished — so report
// (i-1) completed. Claiming 100% while the last step is still working is the
// classic progress-bar lie, and here the last step (blob_extract) is also the
// slowest, so it would sit at "100%" for the longest stretch of the job.
const done = status === "RUNNING" ? step - 1 : step;
return {
step,
total,
name: last[3],
percent: Math.max(0, Math.min(100, Math.floor((done / total) * 100))),
};
}
+23 -5
View File
@@ -100,11 +100,7 @@ export class ReplicationService {
return { ...empty, configured: true, problem: `No se pudo conectar: ${msg}` };
}
const field = (name: string): string | null => {
const m = raw.match(new RegExp(`^\\s*${name}:\\s*(.*)$`, "m"));
const v = m?.[1]?.trim();
return v === undefined || v === "" ? null : v;
};
const field = (name: string): string | null => replicaField(raw, name);
// An empty result set means the server is not configured as a replica at
// all — distinct from "configured but broken", and worth saying plainly.
@@ -151,3 +147,25 @@ export class ReplicationService {
};
}
}
/**
* Read one field out of `SHOW REPLICA STATUS\G` output.
*
* Exported for testing, and worth testing: the obvious regex is wrong.
* `\s` matches newlines in JavaScript, so `^\s*NAME:\s*(.*)$` lets the `\s*`
* after the colon swallow the line break of an EMPTY field and capture the
* following line instead. Last_SQL_Error is empty on a healthy replica, so that
* version reported the next line ("Replicate_Ignore_Server_Ids:") as a SQL
* error and rendered a perfectly healthy replica as broken.
*
* Hence `[^\S\n]` — horizontal whitespace only — on both sides of the name.
*
* @returns the trimmed value, or null when the field is absent OR empty. Empty
* and absent mean the same thing to every caller here: MySQL prints
* error fields as blank rather than omitting them.
*/
export function replicaField(raw: string, name: string): string | null {
const m = raw.match(new RegExp(`^[^\\S\\n]*${name}:[^\\S\\n]*(.*)$`, "m"));
const v = m?.[1]?.trim();
return v === undefined || v === "" ? null : v;
}
+87
View File
@@ -0,0 +1,87 @@
import { replicaField } from "./replication.service";
/**
* Verbatim shape of `SHOW REPLICA STATUS\G` from the live replica, trimmed to
* the fields the panel reads plus the neighbours that matter.
*
* The empty `Last_SQL_Error:` immediately followed by
* `Replicate_Ignore_Server_Ids:` is the whole point of the fixture — that exact
* adjacency is what the first implementation misread.
*/
const HEALTHY = [
"*************************** 1. row ***************************",
" Replica_IO_State: Waiting for source to send event",
" Source_Host: 100.103.77.46",
" Source_User: repl",
" Replica_IO_Running: Yes",
" Replica_SQL_Running: Yes",
" Replicate_Do_DB: ",
" Last_Errno: 0",
" Last_Error: ",
" Seconds_Behind_Source: 0",
" Last_IO_Errno: 0",
" Last_IO_Error: ",
" Last_SQL_Errno: 0",
" Last_SQL_Error: ",
" Replicate_Ignore_Server_Ids: ",
" Source_Server_Id: 1",
].join("\n");
const BROKEN = [
" Replica_IO_Running: Yes",
" Replica_SQL_Running: No",
" Seconds_Behind_Source: NULL",
" Last_IO_Error: ",
" Last_SQL_Error: Could not execute Write_rows event on table jorgecuadros.customers",
" Replicate_Ignore_Server_Ids: ",
].join("\n");
describe("replicaField", () => {
it("reads plain values", () => {
expect(replicaField(HEALTHY, "Replica_IO_Running")).toBe("Yes");
expect(replicaField(HEALTHY, "Replica_SQL_Running")).toBe("Yes");
expect(replicaField(HEALTHY, "Source_Host")).toBe("100.103.77.46");
expect(replicaField(HEALTHY, "Seconds_Behind_Source")).toBe("0");
});
/**
* The regression this file exists for. `\s` matches newlines in JavaScript,
* so `^\s*NAME:\s*(.*)$` walks past an empty field's line break and captures
* the NEXT line — turning a healthy replica into
* "Error SQL: Replicate_Ignore_Server_Ids:" in the admin panel.
*/
it("returns null for an empty field instead of the following line", () => {
expect(replicaField(HEALTHY, "Last_SQL_Error")).toBeNull();
expect(replicaField(HEALTHY, "Last_IO_Error")).toBeNull();
expect(replicaField(HEALTHY, "Last_Error")).toBeNull();
expect(replicaField(HEALTHY, "Replicate_Do_DB")).toBeNull();
expect(replicaField(HEALTHY, "Replicate_Ignore_Server_Ids")).toBeNull();
});
it("still reads a real error when there is one", () => {
expect(replicaField(BROKEN, "Last_SQL_Error")).toBe(
"Could not execute Write_rows event on table jorgecuadros.customers",
);
expect(replicaField(BROKEN, "Replica_SQL_Running")).toBe("No");
});
/** NULL is a distinct state from empty and must survive as the literal. */
it("preserves the literal NULL that MySQL prints for unknown lag", () => {
expect(replicaField(BROKEN, "Seconds_Behind_Source")).toBe("NULL");
});
it("returns null for a field that is not present at all", () => {
expect(replicaField(HEALTHY, "Nonexistent_Field")).toBeNull();
});
/**
* Field names are matched at the start of a line. Without the line anchor,
* "Last_Error" would also match inside "Last_SQL_Error" and read the wrong
* value — the two carry different things and both feed the panel.
*/
it("does not match a field name that is a suffix of another", () => {
const raw = " Last_SQL_Error: boom\n Last_Error: ";
expect(replicaField(raw, "Last_Error")).toBeNull();
expect(replicaField(raw, "Last_SQL_Error")).toBe("boom");
});
});
+1 -1
View File
@@ -1,6 +1,6 @@
{
"name": "@jorgecuadros/web",
"version": "1.0.9",
"version": "1.0.11",
"private": true,
"scripts": {
"dev": "next dev -p 4500",
+68
View File
@@ -225,6 +225,7 @@ function Operaciones() {
</button>
)}
</div>
<JobProgressBar job={activeJob} />
<pre className="ops-log">{activeJob.log || "Iniciando…"}</pre>
</div>
)}
@@ -699,3 +700,70 @@ function KV({ label, value }: { label: string; value: string | null | undefined
</div>
);
}
/**
* Step progress for a running migration.
*
* Only REIMPORT and SYNC report steps; BACKUP and RESTORE are a single
* mysqldump, so they render nothing here rather than a made-up bar — the
* spinner in the heading already says "working".
*
* The safety backup runs before the migration, so `progress` is null for the
* first stretch of every REIMPORT. That phase is named explicitly instead of
* showing 0%, which would read as "stuck".
*/
function JobProgressBar({ job }: { job: OpsJob }) {
const running = job.status === "RUNNING";
const p = job.progress;
if (!p) {
if (!running) return null;
return (
<p className="inline-form-note" style={{ marginTop: 8 }}>
Respaldo de seguridad previo
</p>
);
}
return (
<div style={{ marginTop: 10, marginBottom: 4 }}>
<div
className="row-actions"
style={{ justifyContent: "space-between", marginBottom: 6 }}
>
<span className="inline-form-note" style={{ margin: 0 }}>
Paso {p.step} de {p.total} {p.name}
</span>
<span className="inline-form-note" style={{ margin: 0 }}>
{p.percent}%
</span>
</div>
<div
role="progressbar"
aria-valuenow={p.percent}
aria-valuemin={0}
aria-valuemax={100}
aria-label={`Paso ${p.step} de ${p.total}`}
style={{
height: 6,
borderRadius: 999,
background: "var(--line)",
overflow: "hidden",
}}
>
<div
style={{
width: `${p.percent}%`,
height: "100%",
borderRadius: 999,
transition: "width 400ms ease",
background:
job.status === "FAILED"
? "var(--negative)"
: "var(--positive)",
}}
/>
</div>
</div>
);
}
+10
View File
@@ -61,6 +61,14 @@ export interface UserRow {
export type OpsJobKind = "BACKUP" | "RESTORE" | "REIMPORT" | "SYNC";
export type OpsJobStatus = "RUNNING" | "SUCCESS" | "FAILED";
/** Derived from the job log by the API; null for jobs with no step markers. */
export interface JobProgress {
step: number;
total: number;
name: string;
percent: number;
}
export interface OpsJob {
id: string;
kind: OpsJobKind;
@@ -70,6 +78,8 @@ export interface OpsJob {
createdById: string | null;
startedAt: string;
finishedAt: string | null;
/** Only present on getOpsJob (the polled endpoint), not on the list. */
progress?: JobProgress | null;
}
/**
+10 -3
View File
@@ -78,7 +78,13 @@ SYNC_STEPS = [
]
def run(cmd: list[str]) -> None:
def run(cmd: list[str], step: int | None = None, total: int | None = None) -> None:
# The "[paso i/N] name" marker is a contract with the Operaciones screen,
# which parses the last one to show progress. Emitting it here rather than
# letting the UI count STEPS itself keeps the two from drifting when a step
# is added — the number of steps is only ever stated in this file.
if step is not None and total is not None:
print(f"[paso {step}/{total}] {Path(cmd[1]).name}", flush=True)
print("+ " + " ".join(cmd), flush=True)
r = subprocess.run(cmd)
if r.returncode:
@@ -97,11 +103,12 @@ def main() -> None:
if args.stage:
run([PY, str(HERE / "load_staging.py"), "--output-dir", str(HERE / "output")])
for step in SYNC_STEPS if args.sync else STEPS:
steps = SYNC_STEPS if args.sync else STEPS
for i, step in enumerate(steps, start=1):
cmd = [PY, str(HERE / step), "--env", args.env]
if args.sync:
cmd.append("--sync")
run(cmd)
run(cmd, step=i, total=len(steps))
print(f"\n✓ migration complete for env={args.env}")
+1 -1
View File
@@ -1,6 +1,6 @@
{
"name": "jorgecuadros-platform",
"version": "1.0.9",
"version": "1.0.11",
"private": true,
"workspaces": [
"apps/*",
+1 -1
View File
@@ -1,6 +1,6 @@
{
"name": "@jorgecuadros/database",
"version": "1.0.9",
"version": "1.0.11",
"private": true,
"main": "generated/client/index.js",
"types": "generated/client/index.d.ts",