tests: stabilize test_storage_controller_node_deletion (#10420)

## Problem

`test_storage_controller_node_deletion` sometimes failed because shards
were moving around during timeline creation, and neon_local isn't
tolerant of that. The movements were unexpected because the shards had
only just been created.

This was a regression from #9916

Closes: #10383 

## Summary of changes

- Make this test use multiple AZs -- this makes the storage controller's
scheduling reliably stable

Why this works: in #9916 , I made a simplifying assumption that we would
have multiple AZs to get nice stable scheduling -- it's much easier,
because each tenant has a well defined primary+secondary location when
they have an AZ preference and nodes have different AZs. Everything
still works if you don't have multiple AZs, but you just have this quirk
that sometimes the optimizer can disagree with initial scheduling, so
once in a while a shard moves after being created -- annoying for tests,
harmless IRL.

test_runner/fixtures/neon_fixtures.py

@@ -370,6 +370,7 @@ class NeonEnvBuilder:
         pageserver_config_override: str | Callable[[dict[str, Any]], None] | None = None,
         num_safekeepers: int = 1,
         num_pageservers: int = 1,
+        num_azs: int = 1,
         # Use non-standard SK ids to check for various parsing bugs
         safekeepers_id_start: int = 0,
         # fsync is disabled by default to make the tests go faster
@@ -401,6 +402,7 @@ class NeonEnvBuilder:
         self.pageserver_config_override = pageserver_config_override
         self.num_safekeepers = num_safekeepers
         self.num_pageservers = num_pageservers
+        self.num_azs = num_azs
         self.safekeepers_id_start = safekeepers_id_start
         self.safekeepers_enable_fsync = safekeepers_enable_fsync
         self.auth_enabled = auth_enabled
@@ -990,6 +992,7 @@ class NeonEnv:
         self.endpoints = EndpointFactory(self)
         self.safekeepers: list[Safekeeper] = []
         self.pageservers: list[NeonPageserver] = []
+        self.num_azs = config.num_azs
         self.broker = NeonBroker(self)
         self.pageserver_remote_storage = config.pageserver_remote_storage
         self.safekeepers_remote_storage = config.safekeepers_remote_storage
@@ -1090,14 +1093,21 @@ class NeonEnv:
                 http=self.port_distributor.get_port(),
             )
 
+            # Availabilty zones may also be configured manually with `NeonEnvBuilder.pageserver_config_override`
+            if self.num_azs > 1:
+                # Round-robin assignment of AZ names like us-east-2a, us-east-2b, etc.
+                az_prefix = DEFAULT_AZ_ID[:-1]
+                availability_zone = f"{az_prefix}{chr(ord('a') + (ps_id - 1) % self.num_azs)}"
+            else:
+                availability_zone = DEFAULT_AZ_ID
+
             ps_cfg: dict[str, Any] = {
                 "id": ps_id,
                 "listen_pg_addr": f"localhost:{pageserver_port.pg}",
                 "listen_http_addr": f"localhost:{pageserver_port.http}",
                 "pg_auth_type": pg_auth_type,
                 "http_auth_type": http_auth_type,
-                # Default which can be overriden with `NeonEnvBuilder.pageserver_config_override`
-                "availability_zone": DEFAULT_AZ_ID,
+                "availability_zone": availability_zone,
                 # Disable pageserver disk syncs in tests: when running tests concurrently, this avoids
                 # the pageserver taking a long time to start up due to syncfs flushing other tests' data
                 "no_sync": True,

test_runner/regress/test_storage_controller.py

@@ -2394,6 +2394,7 @@ def test_storage_controller_node_deletion(
     Test that deleting a node works & properly reschedules everything that was on the node.
     """
     neon_env_builder.num_pageservers = 3
+    neon_env_builder.num_azs = 3
     env = neon_env_builder.init_configs()
     env.start()
 
@@ -2407,6 +2408,9 @@ def test_storage_controller_node_deletion(
             tid, placement_policy='{"Attached":1}', shard_count=shard_count_per_tenant
         )
 
+    # Sanity check: initial creations should not leave the system in an unstable scheduling state
+    assert env.storage_controller.reconcile_all() == 0
+
     victim = env.pageservers[-1]
 
     # The procedure a human would follow is: