hippogrief/pobsync
1
0
Fork 0
Code Issues 21 Pull Requests Actions Packages Projects Releases 3 Wiki Activity

(bugfix) Track rsync process state for running backups

Browse Source 
Record rsync process pid and execution phase while normal backup runs are
active so the worker can reconcile stale running rows when rsync has
already disappeared.

Keep finalizing runs out of the missing-process path to avoid marking
slow post-rsync stats collection as a failed transfer.

Closes #54
This commit is contained in:
Peter van Arkel
2026-05-23 00:26:22 +02:00
parent 6eb1b4add3
commit 28da9c4096
5 changed files with 146 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
src/pobsync_backend/tests/test_backup_worker.py
View File

@@ -160,16 +160,19 @@ class BackupWorkerTests(TestCase):
kwargs["state_callback"](
{
"status": "running",
"phase": "rsync",
"snapshot": str(snapshot_dir),
"log": str(log_path),
"rsync": {"command": ["rsync"], "exit_code": None},
"rsync": {"command": ["rsync"], "exit_code": None, "pid": 1234, "pgid": 1234},
}
)
run.refresh_from_db()
self.assertEqual(run.snapshot_path, str(snapshot_dir))
self.assertEqual(run.result["execution"]["phase"], "rsync")
self.assertEqual(run.result["execution"]["log"], str(log_path))
self.assertEqual(run.result["execution"]["snapshot"], str(snapshot_dir))
self.assertEqual(run.result["rsync"]["command"], ["rsync"])
self.assertEqual(run.result["rsync"]["pid"], 1234)
return {
"ok": True,
"dry_run": False,
@@ -228,6 +231,49 @@ class BackupWorkerTests(TestCase):
self.assertEqual(run.result["failure"]["category"], "transport")
self.assertIn("Broken pipe", "\n".join(run.result["rsync"]["log_tail"]))
def test_worker_reconciles_real_run_when_rsync_process_disappears(self) -> None:
with TemporaryDirectory() as tmp:
host = HostConfig.objects.create(host="web-01", address="web-01.example.test")
run = queue_backup_run(host=host)
log_path = Path(tmp) / "backups" / host.host / ".incomplete" / "20260519-021500Z__ABCDEFGH" / "meta" / "rsync.log"
log_path.parent.mkdir(parents=True, exist_ok=True)
log_path.write_text("sending incremental file list\n", encoding="utf-8")
run.status = BackupRun.Status.RUNNING
run.started_at = timezone.now() - timedelta(minutes=10)
run.result["execution"] = {
"phase": "rsync",
"log": str(log_path),
"heartbeat_at": (timezone.now() - timedelta(minutes=10)).isoformat(),
}
run.result["rsync"] = {"pid": 999999, "pgid": 999999, "command": ["rsync"]}
run.save(update_fields=["status", "started_at", "result"])
reconciled = reconcile_running_runs(grace_seconds=300, stale_worker_seconds=24 * 60 * 60)
self.assertEqual(reconciled, 1)
run.refresh_from_db()
self.assertEqual(run.status, BackupRun.Status.FAILED)
self.assertEqual(run.result["failure"]["category"], "rsync_process")
self.assertEqual(run.rsync_exit_code, 255)
def test_worker_does_not_reconcile_missing_rsync_process_during_finalizing_phase(self) -> None:
host = HostConfig.objects.create(host="web-01", address="web-01.example.test")
run = queue_backup_run(host=host)
run.status = BackupRun.Status.RUNNING
run.started_at = timezone.now() - timedelta(minutes=10)
run.result["execution"] = {
"phase": "finalizing",
"heartbeat_at": (timezone.now() - timedelta(minutes=10)).isoformat(),
}
run.result["rsync"] = {"pid": 999999, "pgid": 999999, "command": ["rsync"], "exit_code": 0}
run.save(update_fields=["status", "started_at", "result"])
reconciled = reconcile_running_runs(grace_seconds=300, stale_worker_seconds=24 * 60 * 60)
self.assertEqual(reconciled, 0)
run.refresh_from_db()
self.assertEqual(run.status, BackupRun.Status.RUNNING)
def test_worker_does_not_fail_real_run_for_vanished_file_warning_log(self) -> None:
with TemporaryDirectory() as tmp:
host = HostConfig.objects.create(host="web-01", address="web-01.example.test")