Tests: Test trasformation SSSD does not crash in nss responder after netgroup timeout when backend is offline

SSSD does not crash in nss responder after netgroup timeout when backend is offline

Author: aborah-sudo

src/tests/system/tests/test_netgroups.py

@@ -6,6 +6,9 @@
 
 from __future__ import annotations
 
+import time
+from datetime import datetime as D_T
+
 import pytest
 from sssd_test_framework.roles.ad import AD
 from sssd_test_framework.roles.client import Client
@@ -305,7 +308,66 @@ def test_netgroup__uid_gt_2147483647(client: Client, provider: GenericProvider):
         result = client.tools.getent.passwd(username)
         assert result is not None, f"getent passwd for user '{username}' is empty!"
         assert result.name == username, f"User name '{username}' did not match result '{result.name}'!"
+
     for grpname in ["biggroup1", "biggroup2", "biggroup3"]:
         result = client.tools.getent.group(grpname)
         assert result is not None, f"getent group for group '{grpname}' is empty!"
         assert result.name == grpname, f"Group name '{grpname}' did not match result '{result.name}'!"
+
+
+@pytest.mark.importance("low")
+@pytest.mark.topology(KnownTopologyGroup.AnyProvider)
+def test_netgroup__sssd_entry_cache_nowait_percentage(client: Client, provider: GenericProvider):
+    """
+    :title: Test SSSD netgroup caching with 'entry_cache_nowait_percentage'
+    :setup:
+        1. A netgroup is created and member is added to the netgroup
+        2. SSSD is configured with entry_cache_nowait_percentage
+        3. The domain "test" is updated with `entry_cache_timeout` set to 30 seconds
+        4. SSSD is restarted with the `clean=True` option to apply the new configuration
+    :steps:
+        1. Record the initial response time for netgroup query
+        2. Introduce a network delay of 50 seconds
+        3. Execute another query for netgroup
+        4. Record the response time and verify it is faster than the initial query
+        5. Check the SSSD NSS logs for the presence of the log entry:
+            "Performing midpoint cache update of [netgrp_nowait@test]"
+    :expectedresults:
+        1. The query should return the netgroup with the correct name
+        2. Network delayed for 50 seconds
+        3. Query is executed
+        4. The response time for subsequent queries should be faster than the initial query
+        5. The expected log entries are present in the logs
+    :customerscenario: True
+    """
+    if not isinstance(provider, (LDAP, Samba, AD)):
+        pytest.skip("IPA does not support domain in netgroups")
+
+    netgroup_group = provider.netgroup("netgrp_nowait").add()
+    netgroup_group.add_member(host="host1", user="kau10", domain="example.com")
+    client.sssd.nss.update(
+        filter_groups="root", filter_users="root", debug_level="9", entry_cache_nowait_percentage="50"
+    )
+    client.sssd.dom("test").update(entry_cache_timeout="30")
+    client.sssd.restart(clean=True)
+
+    start = D_T.now()
+    result = client.tools.getent.netgroup(netgroup_group.name)
+    assert result is not None, "Could not get netgroup netgrp_nowait"
+    assert result.name == "netgrp_nowait"
+    end = D_T.now()
+    res_time = end - start
+
+    client.tc.add_delay(provider, "50s")
+    time.sleep(16)
+    start = D_T.now()
+    result = client.tools.getent.netgroup("netgrp_nowait")
+    assert result is not None, "Could not get netgroup netgrp_nowait"
+    assert result.name == "netgrp_nowait"
+    end = D_T.now()
+    catch_response = end - start
+    client.tc.remove_delay(provider)
+
+    read_nss = client.fs.read("/var/log/sssd/sssd_nss.log")
+    assert catch_response < res_time, "Test failed as the cache response time is higher."
+    assert "Performing midpoint cache update of [netgrp_nowait@test]" in read_nss