fix: Fix private key getting lost on stack resume failure

Starting with v7.7.1, we are fetching the private key from the
database. In case of a resume failure, we need to keep the key
in place when running the cleanup step after the failed resume
so that the learner will not lose access to the lab altogether.

Changelog.md

@@ -1,3 +1,9 @@
+Unreleased
+-------------------------
+* [Bug fix] Fix private key getting lost after a stack resume failure.
+  Make sure we keep the stack key in place when running cleanup on
+  a stack that failed to resume.
+
 Version 7.7.2 (2023-09-15)
 -------------------------
 * [Bug fix] Fix editing the `stack_key_type` field in Studio; include

hastexo/tasks.py

@@ -266,7 +266,7 @@ def run(self, **kwargs):
                 'error_msg': textwrap.shorten(e.error_msg, width=256),
                 'ip': None,
                 'user': "",
-                'key': "",
+                'key': self.stack_key if e.suspend else '',
                 'password': "",
                 'provider': provider_name
             }

tests/unit/test_tasks.py

@@ -544,6 +544,7 @@ def test_provider_error_on_cleanup_resume(self):
         # Assertions
         self.assertEqual(stack.status, "RESUME_FAILED")
         self.assertEqual(stack.provider, self.providers[1]["name"])
+        self.assertEqual(self.stack_key, stack.key)
         provider.suspend_stack.assert_called()
 
     def test_undefined_capacity(self):
@@ -1058,6 +1059,10 @@ def test_resume_suspended_stack_persistent_operational_error(self):
         # provider should still be unchanged.
         self.assertEqual(stack.provider, self.providers[1]["name"])
 
+        # When resume fails, we should still have the private key
+        # of the stack.
+        self.assertEqual(self.stack_key, stack.key)
+
     def test_resumed_stack_has_no_ip(self):
         # Setup
         provider = self.mock_providers[1]
@@ -1083,6 +1088,7 @@ def test_resumed_stack_has_no_ip(self):
         # Assertions
         self.assertEqual(stack.status, "RESUME_FAILED")
         self.assertEqual(stack.provider, self.providers[1]["name"])
+        self.assertEqual(self.stack_key, stack.key)
         provider.suspend_stack.assert_called()
 
     def test_timeout_resuming_stack(self):
@@ -1108,6 +1114,7 @@ def test_timeout_resuming_stack(self):
         # Assertions
         self.assertEqual(stack.status, "LAUNCH_TIMEOUT")
         self.assertEqual(stack.provider, self.providers[1]["name"])
+        self.assertEqual(self.stack_key, stack.key)
 
     def test_resume_hook_empty(self):
         # Setup
@@ -1280,6 +1287,7 @@ def test_delete_stack_on_create_failed(self):
 
         # Assertions
         self.assertEqual(stack.status, "CREATE_FAILED")
+        self.assertEqual('', stack.key)
         provider.delete_stack.assert_called_with(
             self.stack_name, False
         )
@@ -1304,6 +1312,7 @@ def test_cleanup_timeout_on_create_failed(self):
 
         # Assertions
         self.assertEqual(stack.status, "CREATE_FAILED")
+        self.assertEqual('', stack.key)
         provider.delete_stack.assert_called_with(
             self.stack_name, False
         )