PR #128 looks up the root credential with account_host = 'localhost',
which never matches under the single root@% account model. Widen the
lookup to account_host IN ('%', 'localhost'), consistent with the
compatibility strategy in deploymentCredentialVars.
- CreateTask now accepts only mysql_root_password; drop
mysql_admin_password and the paired root/admin validation,
keeping the >=16-char length and alphanumeric-only checks
- deploymentCredentialVars extracts mysql_root_password only,
remaining compatible with legacy root@localhost records
- Both delivery playbooks drop xinfra_admin provisioning and
the mysql_admin_password assertion, creating 'root'@'%'
as the remote admin account instead
Wrap task creation in a transaction that locks the business line row (SELECT ... FOR UPDATE) so concurrent requests cannot both pass the availability check. A name is considered occupied by in-flight or retained tasks (pending through registering, finished, register_failed, canceling, rollback_pending/rolling_back/rollback_failed) and by active deployment results; it becomes reusable only after rollback or explicit cleanup releases it.
- POST /delivery/tasks/:id/rollback/retry (platform admin) re-launches
the rollback job after a previous cleanup failure
- POST /delivery/tasks/:id/rollback/release (platform admin) releases
bookkeeping after the operator verified the target host manually
- POST /delivery/tasks/:id/clouddm/retry retries only the CloudDM
registration step for an already healthy instance
- task logs now include rollback AWX job stdout and localized
text/classes for the new rollback states
Introduce a compensating workflow that launches the dedicated AWX
rollback template when a deployment cannot be started or fails midway.
- add task states: rollback_pending, rolling_back, rolled_back,
rollback_failed, rollback_acknowledged
- add RollbackJob model to track the compensating AWX run separately
from the deploy run, preserving both job IDs for audit
- hold resource reservations in 'rollback' status until cleanup
succeeds so a failed cleanup cannot be masked by a later delivery
- poll rollback jobs with a 2-minute launch timeout; an unknown launch
result surfaces as a recoverable failure instead of re-launching
- protect finished/register_failed/rolled-back tasks from rollback;
register_failed keeps the healthy instance and its resource usage
- add DELIVERY_ROLLBACK_TEMPLATE_ID config; without it, failures are
marked rollback_failed and require manual cleanup
- use unique pending-<task_id> placeholder for executor job IDs
- Per-host instance limit (DELIVERY_HOST_INSTANCE_LIMIT, default 4)
- Optional target_host to pin a host from the candidate pool
- Regenerate swagger docs
- 8.0 from Ubuntu archive, 8.4 from MySQL official APT repo
- 5.6/5.7 not supported for now; enforce one MySQL series per host
- Sync Go version whitelist with the playbook package map
- Add topology/port/data_disk, DB options and 8 advanced parameters
with whitelist validation, rendered via extra_vars into the playbook
- Hybrid port allocation over pool 13306-13999
- Tighten bounds: memory 2048-65536 MiB, storage 20-2000 GiB
Implement AWXClient with Launch, GetJob, and Cancel methods for
orchestrating Ansible job templates via AWX REST API. Uses Bearer
token auth with 30s timeout. Includes unit tests with mock transport.
Relates-to: #97
Gmarker689
mac
wonder