Compare commits

...

No commits in common. "c8" and "c8s" have entirely different histories.
c8 ... c8s

12 changed files with 330 additions and 3 deletions

12
.gitignore vendored
View File

@ -1 +1,11 @@
SOURCES/sssd-2.9.4.tar.gz
/sssd-2.6.2.tar.gz
/sssd-2.7.0.tar.gz
/sssd-2.7.2.tar.gz
/sssd-2.7.3.tar.gz
/sssd-2.8.1.tar.gz
/sssd-2.8.2.tar.gz
/sssd-2.9.0.tar.gz
/sssd-2.9.1.tar.gz
/sssd-2.9.2.tar.gz
/sssd-2.9.3.tar.gz
/sssd-2.9.4.tar.gz

View File

@ -1 +0,0 @@
574f6cec9ee12dd943e4305286845343ab7bb891 SOURCES/sssd-2.9.4.tar.gz

View File

@ -0,0 +1,306 @@
From 14f32f681a25aac185d72bc6d22a9e3b59dd265a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Pavel=20B=C5=99ezina?= <pbrezina@redhat.com>
Date: Tue, 30 Apr 2024 12:28:53 +0200
Subject: [PATCH] failover: add failover_primary_timeout option
This was previously hardcoded to 31 seconds (hardcoded retry_timout +
1). This may be too short period under some circumstances.
When we retry primary server we drop connection to the backup server and
if the primary server is not yet available (and there are many
unavailable primary servers) we may go through a long timeout cycle
every half minute.
This patch makes the value configurable.
:config: Added `failover_primary_timout` configuration option. This
can be used to configure how often SSSD tries to reconnect to a
primary server after a successful connection to a backup server.
This was previously hardcoded to 31 seconds which is kept as
the default value.
Resolves: https://github.com/SSSD/sssd/issues/7375
Reviewed-by: Alexey Tikhonov <atikhono@redhat.com>
Reviewed-by: Iker Pedrosa <ipedrosa@redhat.com>
(cherry picked from commit e9738e36937e78f80bb2772c48cffbddf39bd5fe)
---
src/config/SSSDConfig/sssdoptions.py | 2 +
src/config/SSSDConfigTest.py | 2 +
src/config/cfg_rules.ini | 1 +
src/config/etc/sssd.api.conf | 1 +
src/man/sssd.conf.5.xml | 19 ++++++++
src/providers/data_provider.h | 1 +
src/providers/data_provider_fo.c | 14 +++++-
src/providers/fail_over.c | 10 +++++
src/providers/fail_over.h | 3 ++
src/tests/system/tests/test_failover.py | 59 +++++++++++++++++++++++++
10 files changed, 110 insertions(+), 2 deletions(-)
create mode 100644 src/tests/system/tests/test_failover.py
diff --git a/src/config/SSSDConfig/sssdoptions.py b/src/config/SSSDConfig/sssdoptions.py
index 0d75e6d82..95b39aa59 100644
--- a/src/config/SSSDConfig/sssdoptions.py
+++ b/src/config/SSSDConfig/sssdoptions.py
@@ -186,6 +186,8 @@ class SSSDOptions(object):
'dns_resolver_op_timeout': _('How long should keep trying to resolve single DNS query (seconds)'),
'dns_resolver_timeout': _('How long to wait for replies from DNS when resolving servers (seconds)'),
'dns_discovery_domain': _('The domain part of service discovery DNS query'),
+ 'failover_primary_timeout': _('How often SSSD tries to reconnect to the primary server after a successful '
+ 'connection to the backup server.'),
'override_gid': _('Override GID value from the identity provider with this value'),
'case_sensitive': _('Treat usernames as case sensitive'),
'entry_cache_user_timeout': _('Entry cache timeout length (seconds)'),
diff --git a/src/config/SSSDConfigTest.py b/src/config/SSSDConfigTest.py
index b160be2b1..f333c35eb 100755
--- a/src/config/SSSDConfigTest.py
+++ b/src/config/SSSDConfigTest.py
@@ -579,6 +579,7 @@ class SSSDConfigTestSSSDDomain(unittest.TestCase):
'dns_resolver_op_timeout',
'dns_resolver_timeout',
'dns_discovery_domain',
+ 'failover_primary_timeout',
'dyndns_update',
'dyndns_ttl',
'dyndns_iface',
@@ -939,6 +940,7 @@ class SSSDConfigTestSSSDDomain(unittest.TestCase):
'dns_resolver_op_timeout',
'dns_resolver_timeout',
'dns_discovery_domain',
+ 'failover_primary_timeout',
'dyndns_update',
'dyndns_ttl',
'dyndns_iface',
diff --git a/src/config/cfg_rules.ini b/src/config/cfg_rules.ini
index 92e87fb18..4c2ea0b87 100644
--- a/src/config/cfg_rules.ini
+++ b/src/config/cfg_rules.ini
@@ -405,6 +405,7 @@ option = dns_resolver_op_timeout
option = dns_resolver_timeout
option = dns_resolver_use_search_list
option = dns_discovery_domain
+option = failover_primary_timeout
option = override_gid
option = case_sensitive
option = override_homedir
diff --git a/src/config/etc/sssd.api.conf b/src/config/etc/sssd.api.conf
index 5ae6aab19..31787c23c 100644
--- a/src/config/etc/sssd.api.conf
+++ b/src/config/etc/sssd.api.conf
@@ -172,6 +172,7 @@ dns_resolver_server_timeout = int, None, false
dns_resolver_op_timeout = int, None, false
dns_resolver_timeout = int, None, false
dns_discovery_domain = str, None, false
+failover_primary_timeout = int, None, false
override_gid = int, None, false
case_sensitive = str, None, false
override_homedir = str, None, false
diff --git a/src/man/sssd.conf.5.xml b/src/man/sssd.conf.5.xml
index 339f21e25..fbb82e357 100644
--- a/src/man/sssd.conf.5.xml
+++ b/src/man/sssd.conf.5.xml
@@ -3773,6 +3773,25 @@ pam_gssapi_indicators_map = sudo:pkinit, sudo-i:pkinit
</listitem>
</varlistentry>
+ <varlistentry>
+ <term>failover_primary_timeout (integer)</term>
+ <listitem>
+ <para>
+ When no primary server is currently available,
+ SSSD fail overs to a backup server. This option
+ defines the amount of time (in seconds) to
+ wait before SSSD tries to reconnect to a primary
+ server again.
+ </para>
+ <para>
+ Note: The minimum value is 31.
+ </para>
+ <para>
+ Default: 31
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term>override_gid (integer)</term>
<listitem>
diff --git a/src/providers/data_provider.h b/src/providers/data_provider.h
index 36a82b84d..def35e491 100644
--- a/src/providers/data_provider.h
+++ b/src/providers/data_provider.h
@@ -267,6 +267,7 @@ enum dp_res_opts {
DP_RES_OPT_RESOLVER_SERVER_TIMEOUT,
DP_RES_OPT_RESOLVER_USE_SEARCH_LIST,
DP_RES_OPT_DNS_DOMAIN,
+ DP_RES_OPT_FAILOVER_PRIMARY_TIMEOUT,
DP_RES_OPTS /* attrs counter */
};
diff --git a/src/providers/data_provider_fo.c b/src/providers/data_provider_fo.c
index b0aed54e9..c23f92e35 100644
--- a/src/providers/data_provider_fo.c
+++ b/src/providers/data_provider_fo.c
@@ -48,10 +48,20 @@ static int be_fo_get_options(struct be_ctx *ctx,
DP_RES_OPT_RESOLVER_TIMEOUT);
opts->use_search_list = dp_opt_get_bool(ctx->be_res->opts,
DP_RES_OPT_RESOLVER_USE_SEARCH_LIST);
+ opts->primary_timeout = dp_opt_get_int(ctx->be_res->opts,
+ DP_RES_OPT_FAILOVER_PRIMARY_TIMEOUT);
+
opts->retry_timeout = 30;
opts->srv_retry_neg_timeout = 15;
opts->family_order = ctx->be_res->family_order;
+ if (opts->primary_timeout <= opts->retry_timeout) {
+ opts->primary_timeout = opts->retry_timeout + 1;
+ DEBUG(SSSDBG_CONF_SETTINGS,
+ "Warning: failover_primary_timeout is too low, using %lu "
+ "seconds instead\n", opts->primary_timeout);
+ }
+
return EOK;
}
@@ -551,7 +561,7 @@ static void be_resolve_server_done(struct tevent_req *subreq)
struct tevent_req);
struct be_resolve_server_state *state = tevent_req_data(req,
struct be_resolve_server_state);
- time_t timeout = fo_get_service_retry_timeout(state->svc->fo_service) + 1;
+ time_t timeout = fo_get_primary_retry_timeout(state->svc->fo_service);
int ret;
ret = be_resolve_server_process(subreq, state, &new_subreq);
@@ -564,7 +574,6 @@ static void be_resolve_server_done(struct tevent_req *subreq)
}
if (!fo_is_server_primary(state->srv)) {
- /* FIXME: make the timeout configurable */
ret = be_primary_server_timeout_activate(state->ctx, state->ev,
state->ctx, state->svc,
timeout);
@@ -871,6 +880,7 @@ static struct dp_option dp_res_default_opts[] = {
{ "dns_resolver_server_timeout", DP_OPT_NUMBER, { .number = 1000 }, NULL_NUMBER },
{ "dns_resolver_use_search_list", DP_OPT_BOOL, BOOL_TRUE, BOOL_TRUE },
{ "dns_discovery_domain", DP_OPT_STRING, NULL_STRING, NULL_STRING },
+ { "failover_primary_timeout", DP_OPT_NUMBER, { .number = 31 }, NULL_NUMBER },
DP_OPTION_TERMINATOR
};
diff --git a/src/providers/fail_over.c b/src/providers/fail_over.c
index 7cb642448..7f94407c5 100644
--- a/src/providers/fail_over.c
+++ b/src/providers/fail_over.c
@@ -158,6 +158,7 @@ fo_context_init(TALLOC_CTX *mem_ctx, struct fo_options *opts)
ctx->opts->srv_retry_neg_timeout = opts->srv_retry_neg_timeout;
ctx->opts->retry_timeout = opts->retry_timeout;
+ ctx->opts->primary_timeout = opts->primary_timeout;
ctx->opts->family_order = opts->family_order;
ctx->opts->service_resolv_timeout = opts->service_resolv_timeout;
ctx->opts->use_search_list = opts->use_search_list;
@@ -1740,6 +1741,15 @@ time_t fo_get_service_retry_timeout(struct fo_service *svc)
return svc->ctx->opts->retry_timeout;
}
+time_t fo_get_primary_retry_timeout(struct fo_service *svc)
+{
+ if (svc == NULL || svc->ctx == NULL || svc->ctx->opts == NULL) {
+ return 0;
+ }
+
+ return svc->ctx->opts->primary_timeout;
+}
+
bool fo_get_use_search_list(struct fo_server *server)
{
if (
diff --git a/src/providers/fail_over.h b/src/providers/fail_over.h
index 36021ad6f..924a09970 100644
--- a/src/providers/fail_over.h
+++ b/src/providers/fail_over.h
@@ -83,6 +83,7 @@ struct fo_server;
struct fo_options {
time_t srv_retry_neg_timeout;
time_t retry_timeout;
+ time_t primary_timeout;
int service_resolv_timeout;
bool use_search_list;
enum restrict_family family_order;
@@ -211,6 +212,8 @@ int fo_is_srv_lookup(struct fo_server *s);
time_t fo_get_service_retry_timeout(struct fo_service *svc);
+time_t fo_get_primary_retry_timeout(struct fo_service *svc);
+
bool fo_get_use_search_list(struct fo_server *server);
void fo_reset_services(struct fo_ctx *fo_ctx);
diff --git a/src/tests/system/tests/test_failover.py b/src/tests/system/tests/test_failover.py
new file mode 100644
index 000000000..565cec9bc
--- /dev/null
+++ b/src/tests/system/tests/test_failover.py
@@ -0,0 +1,59 @@
+"""
+SSSD Failover tests.
+
+:requirement: Failover
+"""
+
+from __future__ import annotations
+
+import pytest
+from sssd_test_framework.roles.client import Client
+from sssd_test_framework.roles.ldap import LDAP
+from sssd_test_framework.topology import KnownTopology
+
+
+@pytest.mark.parametrize("value, expected", [(None, 31), (15, 31), (60, 60)])
+@pytest.mark.importance("low")
+@pytest.mark.ticket(gh=7375, jira="RHEL-17659")
+@pytest.mark.topology(KnownTopology.LDAP)
+def test_failover__retry_primary(client: Client, ldap: LDAP, value: int | None, expected: int):
+ """
+ :title: Primary server reactivation timeout is respected
+ :setup:
+ 1. Create LDAP user "user-1"
+ 2. Set failover_primary_timeout to @value
+ 3. Set ldap_uri to invalid, not working server
+ 4. Set ldap_backup_uri to working server
+ 5. Start SSSD
+ :steps:
+ 1. Lookup user-1
+ 2. Check that SSSD is connected to backup server
+ 3. Find "Primary server reactivation timeout set to @expected seconds" in domain logs
+ :expectedresults:
+ 1. SSSD failover to backup server
+ 2. SSSD is indeed connected to the backup server
+ 3. String is found
+ :customerscenario: True
+ """
+ ldap.user("user-1").add()
+
+ if value is not None:
+ client.sssd.domain["failover_primary_timeout"] = str(value)
+
+ client.sssd.enable_responder("ifp")
+ client.sssd.domain["ldap_uri"] = "ldap://ldap.invalid"
+ client.sssd.domain["ldap_backup_uri"] = f"ldap://{ldap.host.hostname}"
+ client.sssd.start()
+
+ # Lookup user to make sure SSSD did correctly failover to backup server
+ result = client.tools.id("user-1")
+ assert result is not None
+
+ # Check that SSSD is indeed connected to backup server
+ assert client.sssd.default_domain is not None
+ status = client.sssctl.domain_status(client.sssd.default_domain, active=True)
+ assert ldap.host.hostname in status.stdout
+
+ # Check that primary server reactivation timeout was correctly created
+ log = client.fs.read(client.sssd.logs.domain())
+ assert f"Primary server reactivation timeout set to {expected} seconds" in log
--
2.46.0

7
gating.yaml Normal file
View File

@ -0,0 +1,7 @@
# recipients: sssd-qe, ftrivino
--- !Policy
product_versions:
- rhel-8
decision_context: osci_compose_gate
rules:
- !PassingTestCaseRule {test_case_name: idm-ci.brew-build.tier1.functional}

1
sources Normal file
View File

@ -0,0 +1 @@
SHA512 (sssd-2.9.4.tar.gz) = 9546cf074628f32137b16ca0c763988785271124244b645d1e786762e8578f10d983793a29bffcc004b064452fe8d465476a3041688d2f3c11c2751fb5bec3e2

View File

@ -19,7 +19,7 @@
Name: sssd
Version: 2.9.4
Release: 4%{?dist}
Release: 5%{?dist}
Group: Applications/System
Summary: System Security Services Daemon
License: GPLv3+
@ -33,6 +33,7 @@ Patch0003: 0003-sdap-add-naming_context-as-new-member-of-struct-sdap.patch
Patch0004: 0004-pam-fix-SC-auth-with-multiple-certs-and-missing-logi.patch
Patch0005: 0005-ad-gpo-use-hash-to-store-intermediate-results.patch
Patch0006: 0006-ad-refresh-root-domain-when-read-directly.patch
Patch0007: 0007-failover-add-failover_primary_timeout-option.patch
### Downstream Patches ###
@ -1217,6 +1218,9 @@ fi
%systemd_postun_with_restart sssd.service
%changelog
* Mon Sep 09 2024 Anuar Beisembayev <abeisemb@redhat.com> - 2.9.4-5
- Resolves: RHEL-39085 - [RfE] SSSD Failover Enhancements
* Fri May 17 2024 Arun Bansal <arbansal@redhat.com> - 2.9.4-4
- Resolves: RHEL-33957 - ad: refresh root domain when read directly