orchestratord test: Change upgrade order

Author: def-

ci/nightly/pipeline.template.yml

@@ -2364,6 +2364,7 @@ steps:
     steps:
       - id: orchestratord-defaults
         label: "Orchestratord test (defaults from documentation)"
+        artifact_paths: ["mz_debug_*.zip"]
         depends_on: devel-docker-tags
         timeout_in_minutes: 120
         plugins:
@@ -2376,6 +2377,7 @@ steps:
 
       - id: orchestratord-default-properties
         label: "Orchestratord test (defaults for properties)"
+        artifact_paths: ["mz_debug_*.zip"]
         depends_on: devel-docker-tags
         timeout_in_minutes: 120
         plugins:
@@ -2388,6 +2390,7 @@ steps:
 
       - id: orchestratord-individual
         label: "Orchestratord test (individual properties)"
+        artifact_paths: ["mz_debug_*.zip"]
         depends_on: devel-docker-tags
         timeout_in_minutes: 120
         plugins:
@@ -2400,6 +2403,7 @@ steps:
 
       - id: orchestratord-combine
         label: "Orchestratord test (combine properties)"
+        artifact_paths: ["mz_debug_*.zip"]
         depends_on: build-aarch64
         timeout_in_minutes: 120
         plugins:
@@ -2412,6 +2416,7 @@ steps:
 
       - id: orchestratord-upgrade-individual
         label: "Orchestratord test (upgrade, individual props)"
+        artifact_paths: ["mz_debug_*.zip"]
         depends_on: devel-docker-tags
         timeout_in_minutes: 120
         plugins:
@@ -2427,6 +2432,7 @@ steps:
 
       - id: orchestratord-upgrade-combine
         label: "Orchestratord test (upgrade, combine props)"
+        artifact_paths: ["mz_debug_*.zip"]
         depends_on: devel-docker-tags
         timeout_in_minutes: 120
         plugins:
@@ -2442,6 +2448,7 @@ steps:
 
       - id: orchestratord-upgrade-chain-individual
         label: "Orchestratord test (upgrade chain, individual props)"
+        artifact_paths: ["mz_debug_*.zip"]
         depends_on: devel-docker-tags
         timeout_in_minutes: 120
         plugins:
@@ -2457,6 +2464,7 @@ steps:
 
       - id: orchestratord-upgrade-chain-combine
         label: "Orchestratord test (upgrade chain, combine props)"
+        artifact_paths: ["mz_debug_*.zip"]
         depends_on: devel-docker-tags
         timeout_in_minutes: 120
         plugins:

test/orchestratord/mzcompose.py

@@ -41,6 +41,7 @@
 from materialize.mzcompose.services.balancerd import Balancerd
 from materialize.mzcompose.services.clusterd import Clusterd
 from materialize.mzcompose.services.environmentd import Environmentd
+from materialize.mzcompose.services.mz_debug import MzDebug
 from materialize.mzcompose.services.orchestratord import Orchestratord
 from materialize.mzcompose.services.testdrive import Testdrive
 from materialize.util import all_subclasses
@@ -55,9 +56,23 @@
     Environmentd(),
     Clusterd(),
     Balancerd(),
+    MzDebug(),
 ]
 
 
+def run_mz_debug() -> None:
+    spawn.runv(
+        [
+            "./mz-debug",
+            "self-managed",
+            "--k8s-namespace",
+            "materialize-environment",
+            "--mz-instance-name",
+            "12345678-1234-1234-1234-123456789012",
+        ]
+    )
+
+
 def get_tag(tag: str | None = None) -> str:
     # We can't use the mzbuild tag because it has a different fingerprint for
     # environmentd/clusterd/balancerd and the orchestratord depends on them
@@ -505,7 +520,7 @@ def check() -> None:
             )
             expected = f"{image_registry}/environmentd:{self.value}"
             assert (
-                image == expected
+                image == expected or f"ghcr.io/materializeinc/{image}" == expected
             ), f"Expected environmentd image {expected}, but found {image}"
 
         retry(check, 240)
@@ -1070,11 +1085,11 @@ def check_pods() -> None:
 class AuthenticatorKind(Modification):
     @classmethod
     def values(cls, version: MzVersion) -> list[Any]:
-        # Test None, Password (v0.147.7+), and Sasl (v0.147.16+)
+        # Test None, Password (v0.147.7+), and Sasl
         result = ["None"]
         if version >= MzVersion.parse_mz("v0.147.7"):
             result.append("Password")
-        if version >= MzVersion.parse_mz("v0.147.16"):
+        if version >= MzVersion.parse_mz("v26.0.0"):
             result.append("Sasl")
         return result
 
@@ -1100,13 +1115,13 @@ def validate(self, mods: dict[type[Modification], Any]) -> None:
         if self.value == "Password" and version <= MzVersion.parse_mz("v0.147.6"):
             return
 
-        if self.value == "Sasl" and version < MzVersion.parse_mz("v0.147.16"):
+        if self.value == "Sasl" and version < MzVersion.parse_mz("v26.0.0"):
             return
 
         port = (
             6875
             if (version >= MzVersion.parse_mz("v0.147.0") and self.value == "Password")
-            or (version >= MzVersion.parse_mz("v0.147.16") and self.value == "Sasl")
+            or (version >= MzVersion.parse_mz("v26.0.0") and self.value == "Sasl")
             else 6877
         )
         for i in range(120):
@@ -1252,6 +1267,9 @@ def workflow_defaults(c: Composition, parser: WorkflowArgumentParser) -> None:
     )
     args = parser.parse_args()
 
+    c.up(Service("mz-debug", idle=True))
+    c.invoke("cp", "mz-debug:/usr/local/bin/mz-debug", ".")
+
     current_version = get_tag(args.tag)
 
     # Following https://materialize.com/docs/installation/install-on-local-kind/
@@ -1392,9 +1410,6 @@ def workflow_defaults(c: Composition, parser: WorkflowArgumentParser) -> None:
             materialize_setup = list(yaml.load_all(f, Loader=yaml.Loader))
         assert len(materialize_setup) == 3
 
-        print(version)
-        print(current_version)
-        print(version == current_version)
         if version == current_version:
             materialize_setup[2]["spec"][
                 "environmentdImageRef"
@@ -1493,6 +1508,7 @@ def workflow_defaults(c: Composition, parser: WorkflowArgumentParser) -> None:
                 ]
             )
             raise ValueError("Never completed")
+        run_mz_debug()
 
 
 def workflow_default(c: Composition, parser: WorkflowArgumentParser) -> None:
@@ -1534,7 +1550,8 @@ def workflow_default(c: Composition, parser: WorkflowArgumentParser) -> None:
         "0.29.0"
     ), f"kind >= v0.29.0 required, while you are on {kind_version}"
 
-    c.up(Service("testdrive", idle=True))
+    c.up(Service("testdrive", idle=True), Service("mz-debug", idle=True))
+    c.invoke("cp", "mz-debug:/usr/local/bin/mz-debug", ".")
 
     cluster = "kind"
     clusters = spawn.capture(["kind", "get", "clusters"]).strip().split("\n")
@@ -1806,18 +1823,14 @@ def run_scenario(
                 mod.modify(definition)
                 if mod.value in mod.failed_reconciliation_values():
                     expect_fail = True
-        if not initialize:
-            definition["materialize"]["spec"][
-                "rolloutStrategy"
-            ] = "ImmediatelyPromoteCausingDowntime"
-            definition["materialize"]["spec"]["requestRollout"] = str(uuid.uuid4())
-            run(definition, expect_fail)
         if initialize:
             init(definition)
             run(definition, expect_fail)
             initialize = False  # only initialize once
         else:
             upgrade(definition, expect_fail)
+            definition["materialize"]["spec"]["requestRollout"] = str(uuid.uuid4())
+            run(definition, expect_fail)
         mod_dict = {mod.__class__: mod.value for mod in mods}
         for subclass in all_subclasses(Modification):
             if subclass not in mod_dict:
@@ -1830,6 +1843,7 @@ def run_scenario(
             print(
                 f"Reproduce with bin/mzcompose --find orchestratord run default --recreate-cluster --scenario='{scenario_json}'"
             )
+            run_mz_debug()
             raise
 
 
@@ -2007,3 +2021,5 @@ def post_run_check(definition: dict[str, Any], expect_fail: bool) -> None:
         raise ValueError("Never completed")
     # Wait a bit for the status to stabilize
     time.sleep(60)
+
+    run_mz_debug()