recipe-maintainers/cc-ci
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(2w): WC1.1 reconcile rolls back on deploy FAILURE too (not just unhealthy)

Browse Source 
A broken 'latest' can fail abra's converge (deploy_version raises) rather than
deploy-then-be-unhealthy; wrap the upgrade deploy so BOTH paths trigger the
snapshot-restore rollback instead of crashing the reconcile unit.
This commit is contained in:
autonomic-bot
2026-05-29 01:01:25 +01:00
parent 0812132452
commit 07ea951f31
1 changed files with 11 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
runner/warm_reconcile.py
View File

@ -334,8 +334,17 @@ def reconcile(app: str) -> str:
abra.undeploy(domain) abra.undeploy(domain)
warmsnap.snapshot(recipe, domain, version=last_good) warmsnap.snapshot(recipe, domain, version=last_good)
# snapshot requires undeployed; now bring up latest. # snapshot requires undeployed; now bring up latest.
deploy_version(recipe, domain, latest, dt) # A broken "latest" can fail in two ways: deploy_version raises (abra converge times out on a
if wait_healthy(spec): # crash-looping task) OR it deploys but never becomes healthy. BOTH must roll back, so treat a
# deploy exception the same as an unhealthy result.
upgrade_ok = False
try:
deploy_version(recipe, domain, latest, dt)
upgrade_ok = wait_healthy(spec)
except Exception as e: # noqa: BLE001 — a broken release must trigger rollback, not crash the unit
print(f"[{app}] deploy of latest {latest} failed: {e}", flush=True)
upgrade_ok = False
if upgrade_ok:
write_last_good(recipe, latest) write_last_good(recipe, latest)
print(f"[{app}] upgrade healthy → committed last-good={latest}", flush=True) print(f"[{app}] upgrade healthy → committed last-good={latest}", flush=True)
return f"upgraded:{last_good}->{latest}" return f"upgraded:{last_good}->{latest}"