Remove flaky tests (#2129)

Fix #1930

Author: nfx

tests/integration/assessment/test_ext_hms.py

@@ -1,24 +1,20 @@
 import dataclasses
-from datetime import timedelta
 
-from databricks.sdk.errors import (
-    InvalidParameterValue,
-    NotFound,
-)
-from databricks.sdk.retries import retried
+from databricks.labs.lsql.backends import CommandExecutionBackend
 from databricks.sdk.service.iam import PermissionLevel
 
 
-@retried(on=[NotFound, InvalidParameterValue], timeout=timedelta(minutes=5))
 def test_running_real_assessment_job_ext_hms(
     ws,
     installation_ctx,
     env_or_skip,
     make_cluster_policy,
     make_cluster_policy_permissions,
 ):
+    cluster_id = env_or_skip('TEST_EXT_HMS_CLUSTER_ID')
     ext_hms_ctx = installation_ctx.replace(
         skip_dashboards=True,
+        sql_backend=CommandExecutionBackend(ws, cluster_id),
         config_transform=lambda wc: dataclasses.replace(
             wc,
             override_clusters=None,

tests/integration/hive_metastore/test_workflows.py

@@ -1,11 +1,7 @@
-from datetime import timedelta
-
 import pytest
-from databricks.sdk.errors import InvalidParameterValue, NotFound
-from databricks.sdk.retries import retried
+from databricks.sdk.errors import NotFound
 
 
-@retried(on=[NotFound, InvalidParameterValue], timeout=timedelta(minutes=10))
 @pytest.mark.parametrize(
     "prepare_tables_for_migration,workflow",
     [
@@ -63,7 +59,6 @@ def test_table_migration_job_refreshes_migration_status(
     assert len(asserts) == 0, assert_message
 
 
-@retried(on=[NotFound], timeout=timedelta(minutes=8))
 @pytest.mark.parametrize('prepare_tables_for_migration', [('hiveserde')], indirect=True)
 def test_hiveserde_table_in_place_migration_job(ws, installation_ctx, prepare_tables_for_migration):
     tables, dst_schema = prepare_tables_for_migration
@@ -84,7 +79,6 @@ def test_hiveserde_table_in_place_migration_job(ws, installation_ctx, prepare_ta
             assert False, f"{table.name} not found in {dst_schema.catalog_name}.{dst_schema.name}"
 
 
-@retried(on=[NotFound], timeout=timedelta(minutes=8))
 @pytest.mark.parametrize('prepare_tables_for_migration', [('hiveserde')], indirect=True)
 def test_hiveserde_table_ctas_migration_job(ws, installation_ctx, prepare_tables_for_migration):
     tables, dst_schema = prepare_tables_for_migration

tests/integration/install/test_installation.py

@@ -1,14 +1,12 @@
 import dataclasses
 import json
 import logging
-from dataclasses import replace
 from datetime import timedelta
 
 import pytest  # pylint: disable=wrong-import-order
 from databricks.labs.ucx.__about__ import __version__
 
 from databricks.labs.blueprint.installation import Installation
-from databricks.labs.blueprint.installer import RawState
 from databricks.labs.blueprint.parallel import ManyError
 from databricks.labs.blueprint.tui import MockPrompts
 from databricks.labs.blueprint.wheels import ProductInfo
@@ -356,74 +354,6 @@ def test_check_inventory_database_exists(ws, installation_ctx):
         installation_ctx.workspace_installer.configure()
 
 
-@retried(on=[NotFound], timeout=timedelta(minutes=5))
-@pytest.mark.parametrize('prepare_tables_for_migration', [('regular')], indirect=True)
-def test_table_migration_job(ws, installation_ctx, env_or_skip, prepare_tables_for_migration):
-
-    ctx = installation_ctx.replace(
-        config_transform=lambda wc: replace(wc, override_clusters=None),
-        extend_prompts={
-            r"Parallelism for migrating.*": "1000",
-            r"Min workers for auto-scale.*": "2",
-            r"Max workers for auto-scale.*": "20",
-            r"Instance pool id to be set.*": env_or_skip("TEST_INSTANCE_POOL_ID"),
-            r".*Do you want to update the existing installation?.*": 'yes',
-        },
-    )
-    tables, dst_schema = prepare_tables_for_migration
-
-    ctx.workspace_installation.run()
-    ctx.deployed_workflows.run_workflow("migrate-tables")
-    # assert the workflow is successful
-    assert ctx.deployed_workflows.validate_step("migrate-tables")
-    # assert the tables are migrated
-    for table in tables.values():
-        try:
-            assert ws.tables.get(f"{dst_schema.catalog_name}.{dst_schema.name}.{table.name}").name
-        except NotFound:
-            assert False, f"{table.name} not found in {dst_schema.catalog_name}.{dst_schema.name}"
-    # assert the cluster is configured correctly
-    for job_cluster in ws.jobs.get(
-        ctx.installation.load(RawState).resources["jobs"]["migrate-tables"]
-    ).settings.job_clusters:
-        if job_cluster.job_cluster_key != "table_migration":
-            # don't assert on the cluster for parse logs task
-            continue
-        assert job_cluster.new_cluster.autoscale.min_workers == 2
-        assert job_cluster.new_cluster.autoscale.max_workers == 20
-        assert job_cluster.new_cluster.spark_conf["spark.sql.sources.parallelPartitionDiscovery.parallelism"] == "1000"
-
-
-@retried(on=[NotFound], timeout=timedelta(minutes=8))
-@pytest.mark.parametrize('prepare_tables_for_migration', [('regular')], indirect=True)
-def test_table_migration_job_cluster_override(ws, installation_ctx, prepare_tables_for_migration, env_or_skip):
-
-    tables, dst_schema = prepare_tables_for_migration
-    ctx = installation_ctx.replace(
-        extend_prompts={
-            r".*Do you want to update the existing installation?.*": 'yes',
-        },
-    )
-    ctx.workspace_installation.run()
-    ctx.deployed_workflows.run_workflow("migrate-tables")
-    # assert the workflow is successful
-    assert ctx.deployed_workflows.validate_step("migrate-tables")
-    # assert the tables are migrated
-    for table in tables.values():
-        try:
-            assert ws.tables.get(f"{dst_schema.catalog_name}.{dst_schema.name}.{table.name}").name
-        except NotFound:
-            assert False, f"{table.name} not found in {dst_schema.catalog_name}.{dst_schema.name}"
-    # assert the cluster is configured correctly on the migrate tables tasks
-    install_state = ctx.installation.load(RawState)
-    job_id = install_state.resources["jobs"]["migrate-tables"]
-    assert all(
-        task.existing_cluster_id == env_or_skip("TEST_USER_ISOLATION_CLUSTER_ID")
-        for task in ws.jobs.get(job_id).settings.tasks
-        if task.task_key != "parse_logs"
-    )
-
-
 def test_compare_remote_local_install_versions(ws, installation_ctx):
     installation_ctx.workspace_installation.run()
     with pytest.raises(