update backup & HA tests (#189)

### Problem
<!-- What issue is this PR trying to solve? -->
1. keys for AWS changed so tests no longer passed
2. HA test check for all db processes down fails

### Solution
<!-- A summary of the solution addressing the above issue -->
1. update backup configs in backup tests for the new AWS keys/configs 
2. update HA test check for all db processes down to ignore the
mongodb_exporter process

lib/charms/mongodb/v0/mongodb_backups.py

@@ -107,7 +107,7 @@ def _on_s3_credential_changed(self, event: CredentialsChangedEvent):
             return
 
         # pbm requires that the URI is set before adding configs
-        pbm_snap.set({"uri": self._backup_config.uri})
+        pbm_snap.set({"pbm-uri": self._backup_config.uri})
 
         # Add and sync configuration options while handling errors related to configuring options
         # and re-syncing PBM.

lib/charms/operator_libs_linux/v1/snap.py

@@ -83,7 +83,7 @@
 
 # Increment this PATCH version before using `charmcraft publish-lib` or reset
 # to 0 if you are raising the major API version
-LIBPATCH = 8
+LIBPATCH = 9
 
 
 # Regex to locate 7-bit C1 ANSI sequences
@@ -976,19 +976,27 @@ def _system_set(config_item: str, value: str) -> None:
         raise SnapError("Failed setting system config '{}' to '{}'".format(config_item, value))
 
 
-def hold_refresh(days: int = 90) -> bool:
+def hold_refresh(days: int = 90, forever: bool = False) -> bool:
     """Set the system-wide snap refresh hold.
 
     Args:
         days: number of days to hold system refreshes for. Maximum 90. Set to zero to remove hold.
+        forever: if True, will set a hold forever.
     """
-    # Currently the snap daemon can only hold for a maximum of 90 days
-    if not isinstance(days, int) or days > 90:
-        raise ValueError("days must be an int between 1 and 90")
+    if not isinstance(forever, bool):
+        raise TypeError("forever must be a bool")
+    if not isinstance(days, int):
+        raise TypeError("days must be an int")
+    if forever:
+        _system_set("refresh.hold", "forever")
+        logger.info("Set system-wide snap refresh hold to: forever")
     elif days == 0:
         _system_set("refresh.hold", "")
         logger.info("Removed system-wide snap refresh hold")
     else:
+        # Currently the snap daemon can only hold for a maximum of 90 days
+        if not 1 <= days <= 90:
+            raise ValueError("days must be between 1 and 90")
         # Add the number of days to current time
         target_date = datetime.now(timezone.utc).astimezone() + timedelta(days=days)
         # Format for the correct datetime format

tests/integration/backup_tests/test_backups.py

@@ -106,9 +106,10 @@ async def test_ready_correct_conf(ops_test: OpsTest) -> None:
     choices = string.ascii_letters + string.digits
     unique_path = "".join([secrets.choice(choices) for _ in range(4)])
     configuration_parameters = {
-        "bucket": "pbm-test-bucket-1",
+        "bucket": "data-charms-testing",
         "path": f"mongodb-vm/test-{unique_path}",
-        "region": "us-west-2",
+        "endpoint": "https://s3.amazonaws.com",
+        "region": "us-east-1",
     }
 
     # apply new configuration options
@@ -218,9 +219,9 @@ async def test_multi_backup(ops_test: OpsTest, continuous_writes_to_db) -> None:
     # set AWS credentials, set configs for s3 storage, and wait to resync
     await helpers.set_credentials(ops_test, cloud="AWS")
     configuration_parameters = {
-        "bucket": "pbm-test-bucket-1",
-        "region": "us-west-2",
-        "endpoint": "",
+        "bucket": "data-charms-testing",
+        "region": "us-east-1",
+        "endpoint": "https://s3.amazonaws.com",
     }
     await ops_test.model.applications[S3_APP_NAME].set_config(configuration_parameters)
     await asyncio.gather(
@@ -299,9 +300,9 @@ async def test_restore_new_cluster(ops_test: OpsTest, add_writes_to_db, cloud_pr
     await helpers.set_credentials(ops_test, cloud=cloud_provider)
     if cloud_provider == "AWS":
         configuration_parameters = {
-            "bucket": "pbm-test-bucket-1",
-            "region": "us-west-2",
-            "endpoint": "",
+            "bucket": "data-charms-testing",
+            "region": "us-east-1",
+            "endpoint": "https://s3.amazonaws.com",
         }
     else:
         configuration_parameters = {

tests/integration/ha_tests/helpers.py

@@ -34,6 +34,8 @@
 MONGOD_SERVICE_DEFAULT_PATH = "/etc/systemd/system/snap.charmed-mongodb.mongod.service"
 TMP_SERVICE_PATH = "tests/integration/ha_tests/tmp.service"
 LOGGING_OPTIONS = f"--logpath={MONGO_COMMON_DIR}/var/log/mongod/mongodb.log --logappend"
+EXPORTER_PROC = "/usr/bin/mongodb_exporter"
+GREP_PROC = "grep"
 
 
 class ProcessError(Exception):
@@ -553,20 +555,15 @@ async def all_db_processes_down(ops_test: OpsTest) -> bool:
     app = await app_name(ops_test)
 
     try:
-        for attempt in Retrying(stop=stop_after_attempt(50), wait=wait_fixed(3)):
+        for attempt in Retrying(stop=stop_after_attempt(60), wait=wait_fixed(3)):
             with attempt:
                 for unit in ops_test.model.applications[app].units:
-                    search_db_process = f"run --unit {unit.name} ps aux | grep {DB_PROCESS}"
+                    search_db_process = f"run --unit {unit.name} pgrep -x mongod"
                     _, processes, _ = await ops_test.juju(*search_db_process.split())
-
-                    # `ps aux | grep {DB_PROCESS}` is a process on it's own and will be shown in
-                    # the output of ps aux, hence it it is important that we check if there is
-                    # more than one process containing the name `DB_PROCESS`
                     # splitting processes by "\n" results in one or more empty lines, hence we
                     # need to process these lines accordingly.
                     processes = [proc for proc in processes.split("\n") if len(proc) > 0]
-
-                    if len(processes) > 1:
+                    if len(processes) > 0:
                         raise ProcessRunningError
     except RetryError:
         return False