CloudFormation change sets, Aurora Stop/Start, and S3 delete conditions

August 20, 2026 · v1.4.21

v1.4.21 is a CloudFormation and Aurora correctness release. CloudFormation change sets stop reporting an invalid status that broke the CDK, a replacement of a custom-named resource is refused instead of silently destroying data, and SES configuration sets provision. RDS Aurora global databases and replicated readers survive Stop/Start and validate engine versions across every write path. Around that, S3 gains correct conditional deletes and multipart composite checksums, IoT adds a native mTLS MQTT listener on port 8883, and API Gateway routes on the resource+method pair the way AWS does.

New

• IoT Core — native mTLS MQTT on port 8883. The embedded broker spoke MQTT only over WebSocket, so AWS IoT Device SDK binaries could not connect. It now also listens for MQTT over TLS on 8883 — on by default when the cryptography package is installed (IOT_MTLS_ENABLED=0 to disable, IOT_MTLS_PORT to move it) — with the broker certificate served at GET /_ministack/iot/ca.pem. A client certificate is optional: a device presenting none is served under MINISTACK_ACCOUNT_ID like an unsigned WebSocket upgrade, and an unknown or non-ACTIVE certificate is refused with a 0x05 CONNACK. Contributed by @iot-rocket.
• RDS — replicated Aurora PostgreSQL readers survive Stop/Start and warm boot. Under MINISTACK_RDS_PG_CLUSTER_REPLICATION, StopDBCluster now stops the reader containers alongside the writer and StartDBCluster revives each reader by re-cloning from the writer; the same revival runs on a MiniStack restart, so a persisted reader comes back as a real hot standby instead of being demoted to a writer alias. Contributed by @kiran01bm.
• CloudFormation — AWS::SES::ConfigurationSet and AWS::SES::ConfigurationSetEventDestination. A template carrying an SES configuration set failed with Unsupported resource type; both resource types now provision (registered in the classic and v2 SES stores, Ref returns the set name, the event destination round-trips), CloudFormation-provisioning fidelity only. Contributed by @ryan-bennett.

Fixes

• API Gateway (REST) — routing is on the resource+method pair, not the resource alone. A request a matched resource does not serve — a methodless intermediate node, a CORS-preflight-only node, or an undeclared verb — now falls through to a {proxy+} elsewhere in the tree as it does on AWS, and only an exact resource+method match keeps the request; routing precedes authorization. Supersedes the 1.4.16 change that answered a methodless resource 403. Contributed by @iot-rocket.
• CloudFormation — change sets report a valid status, fail when execution fails, and do not outlive their stack. ExecuteChangeSet wrote the invalid EXECUTE_COMPLETE into Status, breaking the CDK's Status == CREATE_COMPLETE gate. Status now stays CREATE_COMPLETE while ExecutionStatus tracks the deployment outcome; a no-change set ends FAILED, a missing set returns ChangeSetNotFound (404), a duplicate name is AlreadyExistsException, deleting a stack removes its change sets, and a direct UpdateStack marks pending sets OBSOLETE. Reported by @iot-rocket.
• CloudFormation — a replacement of a custom-named resource is refused instead of destroying data. An update requiring replacement of a resource with an explicit physical name (for example changing a DynamoDB key attribute's type on a table with a TableName) executed the replacement and dropped the table's rows; it now fails and rolls back to UPDATE_ROLLBACK_COMPLETE with the message real CloudFormation uses, leaving the resource and its data intact. Reported by @iot-rocket.
• RDS — Aurora engine versions are validated on non-create writes. ModifyDBInstance, ModifyDBCluster, CreateGlobalCluster and ModifyGlobalCluster stored engine versions the catalog does not advertise; all paths now reject them (InvalidParameterCombination / Cannot find upgrade target … on the modify paths), ModifyGlobalCluster propagates an accepted version to every member, and a member moved to a different major than its global is refused. Contributed by @kiran01bm.
• RDS — global Aurora stop/start preserves topology and MySQL replication. StopDBCluster and StartDBCluster are now limited to sole-member global databases (InvalidDBClusterStateFault), deleting a primary's last instance preserves compute other global members still need, and a successful recreate resets stale MySQL replica state before re-linking replication. Contributed by @kiran01bm.
• Step Functions — an unimplemented optimized integration fails instead of passing through. A Task using an arn:aws:states:::<service>:<action> integration MiniStack does not implement echoed its input back as SUCCEEDED, so a state machine whose side effect never ran still reported success; it now fails with States.Runtime. Reported by @iwasakar.
• S3 — a conditional delete of an absent key is answered correctly. DeleteObject with If-Match on a key that is not there now returns 204, and If-Match: * is honored as an existence check — it holds against any present object and returns 412 PreconditionFailed when the key is absent — rather than being compared as a literal ETag. Contributed by @gaul.
• S3 — multipart uploads carry their parts' checksums through to a composite. CreateMultipartUpload records the checksum algorithm, UploadPart validates and echoes each part's checksum, and completion builds the AWS composite (<digest>-<parts>, the digest of the parts' digests) read back as COMPOSITE; a mismatch is BadDigest. Contributed by @gaul.

Upgrade

docker pull ministackorg/ministack:1.4.21
docker run -d -p 4566:4566 ministackorg/ministack:1.4.21

Or pin in compose.yaml:

services:
  ministack:
    image: ministackorg/ministack:1.4.21
    ports:
      - "4566:4566"