Propagate health status from SB Service_Monitor to NB Logical_Switch_Port_Health_Check table.
This synchronization enables CMS (Cloud Management System) to make informed decisions about virtual machine availability, routing, and failover based on the current health status of monitored service Signed-off-by: Alexandra Rukomoinikova <[email protected]> --- v1 --> v2: added DOC --- northd/en-sync-from-sb.c | 4 ++ northd/en-sync-from-sb.h | 3 ++ northd/inc-proc-northd.c | 8 +++ northd/northd.c | 83 ++++++++++++++++++++++++++++++++ ovn-nb.ovsschema | 8 ++- ovn-nb.xml | 8 +++ tests/ovn-inc-proc-graph-dump.at | 1 + tests/system-ovn.at | 6 +++ 8 files changed, 119 insertions(+), 2 deletions(-) diff --git a/northd/en-sync-from-sb.c b/northd/en-sync-from-sb.c index 6d4ff3e39..11f3886dc 100644 --- a/northd/en-sync-from-sb.c +++ b/northd/en-sync-from-sb.c @@ -52,6 +52,10 @@ en_sync_from_sb_get_input_data(struct engine_node *node, EN_OVSDB_GET(engine_get_input("SB_port_binding", node)); data->sb_ha_ch_grp_table = EN_OVSDB_GET(engine_get_input("SB_ha_chassis_group", node)); + data->sbrec_service_monitor_by_service_type = + engine_ovsdb_node_get_index( + engine_get_input("SB_service_monitor", node), + "sbrec_service_monitor_by_service_type"); } enum engine_node_state diff --git a/northd/en-sync-from-sb.h b/northd/en-sync-from-sb.h index bea248c45..177d75de0 100644 --- a/northd/en-sync-from-sb.h +++ b/northd/en-sync-from-sb.h @@ -7,6 +7,9 @@ struct en_sync_from_sb_data { /* Southbound table references */ const struct sbrec_port_binding_table *sb_pb_table; const struct sbrec_ha_chassis_group_table *sb_ha_ch_grp_table; + + /* Indexes */ + struct ovsdb_idl_index *sbrec_service_monitor_by_service_type; }; void *en_sync_from_sb_init(struct engine_node *, struct engine_arg *); diff --git a/northd/inc-proc-northd.c b/northd/inc-proc-northd.c index 732066638..8f6f491fb 100644 --- a/northd/inc-proc-northd.c +++ b/northd/inc-proc-northd.c @@ -478,6 +478,7 @@ void inc_proc_northd_init(struct ovsdb_idl_loop *nb, sync_from_sb_northd_handler); engine_add_input(&en_sync_from_sb, &en_sb_port_binding, NULL); engine_add_input(&en_sync_from_sb, &en_sb_ha_chassis_group, NULL); + engine_add_input(&en_sync_from_sb, &en_sb_service_monitor, NULL); engine_add_input(&en_northd_output, &en_acl_id, NULL); engine_add_input(&en_northd_output, &en_sync_from_sb, NULL); @@ -593,6 +594,13 @@ void inc_proc_northd_init(struct ovsdb_idl_loop *nb, "sbrec_service_monitor_by_learned_type", sbrec_service_monitor_by_learned_type); + struct ovsdb_idl_index *sbrec_service_monitor_by_service_type + = ovsdb_idl_index_create1(sb->idl, + &sbrec_service_monitor_col_type); + engine_ovsdb_node_add_index(&en_sb_service_monitor, + "sbrec_service_monitor_by_service_type", + sbrec_service_monitor_by_service_type); + struct ed_type_global_config *global_config = engine_get_internal_data(&en_global_config); unixctl_command_register("debug/chassis-features-list", "", 0, 0, diff --git a/northd/northd.c b/northd/northd.c index 1904b162b..b11a57ca9 100644 --- a/northd/northd.c +++ b/northd/northd.c @@ -21089,6 +21089,85 @@ handle_port_binding_changes( hmapx_destroy(&lr_groups); } +static bool +svc_monitor_matches_health_check(const struct sbrec_service_monitor *sbrec_mon, + const char *protocol, + const char *src_ip, + const char *target_ip, + uint16_t destination_port) +{ + if (strcmp(sbrec_mon->protocol, protocol) != 0) { + return false; + } + + if (!strcmp(protocol, "tcp") || !strcmp(protocol, "udp")) { + if (sbrec_mon->port != destination_port) { + return false; + } + } + + if (!strcmp(sbrec_mon->src_ip, src_ip)) { + return true; + } + + if (!strcmp(sbrec_mon->ip, target_ip)) { + return true; + } + + return false; +} + +static void +handle_service_monitor_changes( + struct ovsdb_idl_index *sbrec_service_monitor_by_service_type, + const struct hmap *ls_ports) +{ + const struct sbrec_service_monitor *sbrec_mon; + struct sbrec_service_monitor *key = + sbrec_service_monitor_index_init_row( + sbrec_service_monitor_by_service_type); + + sbrec_service_monitor_set_type(key, "logical-switch-port"); + + SBREC_SERVICE_MONITOR_FOR_EACH_EQUAL (sbrec_mon, key, + sbrec_service_monitor_by_service_type) { + + struct ovn_port *op = ovn_port_find(ls_ports, + sbrec_mon->logical_port); + if (!op) { + continue; + } + + ovs_assert(op->nbsp && op->nbsp->n_health_checks); + + /* There shouldn't be many health checks on port, + * liniar check shouldn't be heavy. */ + for (size_t i = 0; i < op->nbsp->n_health_checks; i++) { + const struct nbrec_logical_switch_port_health_check *lsp_hc = + op->nbsp->health_checks[i]; + + if (!svc_monitor_matches_health_check(sbrec_mon, + lsp_hc->protocol, + lsp_hc->src_ip, + lsp_hc->address, + lsp_hc->port)) { + continue; + } + + const char *desired_status = sbrec_mon->status; + if (desired_status) { + if (!lsp_hc->status || + strcmp(lsp_hc->status, desired_status)) { + nbrec_logical_switch_port_health_check_set_status( + lsp_hc, sbrec_mon->status); + } + } + } + } + + sbrec_service_monitor_index_destroy_row(key); +} + /* Handle a fairly small set of changes in the southbound database. */ void ovnsb_db_run(struct ovsdb_idl_txn *ovnsb_txn, @@ -21106,6 +21185,10 @@ ovnsb_db_run(struct ovsdb_idl_txn *ovnsb_txn, &northd_data->lr_ports, &ha_ref_chassis_map); + handle_service_monitor_changes( + sync_from_sb_data->sbrec_service_monitor_by_service_type, + &northd_data->ls_ports); + update_sb_ha_group_ref_chassis(sync_from_sb_data->sb_ha_ch_grp_table, &ha_ref_chassis_map); diff --git a/ovn-nb.ovsschema b/ovn-nb.ovsschema index 538b37456..6892713dc 100644 --- a/ovn-nb.ovsschema +++ b/ovn-nb.ovsschema @@ -1,7 +1,7 @@ { "name": "OVN_Northbound", - "version": "7.16.0", - "cksum": "3492890733 44939", + "version": "7.17.0", + "cksum": "4065801885 45152", "tables": { "NB_Global": { "columns": { @@ -264,6 +264,10 @@ "address": {"type": {"key": "string", "min": 0, "max": 1}}, + "status": { + "type": {"key": {"type": "string", + "enum": ["set", ["online", "offline", "error"]]}, + "min": 0, "max": 1}}, "options": { "type": {"key": "string", "value": "string", diff --git a/ovn-nb.xml b/ovn-nb.xml index e60edbd8d..e71d33b2a 100644 --- a/ovn-nb.xml +++ b/ovn-nb.xml @@ -2225,6 +2225,14 @@ IP address to monitor for the health check. </column> + <column name="status"> + Health status synchronized from the <code>status</code> field of + corresponding service monitor in the SBDB. This status indicates + the current availability of logical switch port for the CMS to + determine service virtual machine health and make routing or + failover decisions. + </column> + <column name="options" key="interval" type='{"type": "integer"}'> The interval, in seconds, between service monitor checks. </column> diff --git a/tests/ovn-inc-proc-graph-dump.at b/tests/ovn-inc-proc-graph-dump.at index ff2c8c0c7..0122a6a4f 100644 --- a/tests/ovn-inc-proc-graph-dump.at +++ b/tests/ovn-inc-proc-graph-dump.at @@ -99,6 +99,7 @@ digraph "Incremental-Processing-Engine" { northd -> sync_from_sb [[label="sync_from_sb_northd_handler"]]; SB_port_binding -> sync_from_sb [[label=""]]; SB_ha_chassis_group -> sync_from_sb [[label=""]]; + SB_service_monitor -> sync_from_sb [[label=""]]; lr_nat [[style=filled, shape=box, fillcolor=white, label="lr_nat"]]; northd -> lr_nat [[label="lr_nat_northd_handler"]]; lr_stateful [[style=filled, shape=box, fillcolor=white, label="lr_stateful"]]; diff --git a/tests/system-ovn.at b/tests/system-ovn.at index 14c0996a0..3ac4f2442 100644 --- a/tests/system-ovn.at +++ b/tests/system-ovn.at @@ -21033,6 +21033,7 @@ check_row_count sb:Service_Monitor 1 # Wait until the services are set to online. wait_row_count Service_Monitor 1 status=online +wait_row_count nb:Logical_Switch_Port_Health_Check 1 status=online protocol=icmp check ovn-nbctl lsp-hc-add lport tcp 192.168.0.255 4041 192.168.0.10 @@ -21043,12 +21044,14 @@ NETNS_DAEMONIZE([lport], [nc -l -k 192.168.0.10 4041], [lport_tcp.pid]) # Wait until the services are set to online. wait_row_count Service_Monitor 2 status=online +wait_row_count nb:Logical_Switch_Port_Health_Check 1 status=online protocol=tcp check ovn-nbctl lsp-hc-add lport udp 192.168.0.255 4042 192.168.0.10 NETNS_DAEMONIZE([lport], [nc -ulp 4042], [lport_udp.pid]) # Wait until the services are set to online. +wait_row_count nb:Logical_Switch_Port_Health_Check 1 status=online protocol=udp wait_row_count Service_Monitor 3 status=online check ovn-nbctl lsp-hc-del lport @@ -21059,6 +21062,7 @@ check_row_count sb:Service_Monitor 1 # Wait until the services are set to online. wait_row_count Service_Monitor 1 status=online +wait_row_count nb:Logical_Switch_Port_Health_Check 1 status=online protocol=icmp # IPv6 TCP health check check ovn-nbctl lsp-hc-add lport tcp 2001:db8::ff 4043 2001:db8::10 @@ -21070,6 +21074,7 @@ NETNS_DAEMONIZE([lport], [nc -6 -l -k 2001:db8::10 4043], [lport_ipv6_tcp.pid]) # Wait until the services are set to online. wait_row_count Service_Monitor 2 status=online +wait_row_count nb:Logical_Switch_Port_Health_Check 1 status=online protocol=tcp # IPv6 UDP health check check ovn-nbctl lsp-hc-add lport udp 2001:db8::ff 4044 2001:db8::10 @@ -21080,6 +21085,7 @@ NETNS_DAEMONIZE([lport], [nc -6 -u -l 2001:db8::10 4044], [lport_ipv6_udp.pid]) # Wait until the services are set to online. wait_row_count Service_Monitor 3 status=online +wait_row_count nb:Logical_Switch_Port_Health_Check 1 status=online protocol=udp killall nc 2>/dev/null || true -- 2.48.1 _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
