name value info need_reboot visible_level enable_connection_diagnosis True if enable, obproxy will record disconenction info false USER read_stale_remove_interval 6h read stale feedback remove interval false USER ob_max_read_stale_time -1 max read stale time, unit us, -1 means close server read stale check false USER read_stale_retry_interval 5s read stale retry interval false USER enable_session_info_verification False if enabled, obproxy will send last server session info to the observer which is different to the last server session for session verification false USER init_sql proxy will send init sql to observer after login success false SYS enable_binlog_service True if enabled, obproxy will send binlog request to OBLogProxy false SYS binlog_service_ip 10.252.210.147:2983 binlog service ip, format ip1:sql_port1 false SYS weak_read_user_list weak read for list of users, format user1;user2 false USER enable_transaction_split False if enabled, support transaction split false SYS enable_read_write_split False if enabled, use read write split mode false SYS enable_global_ps_cache True if enabled, use global ps cache true SYS obproxy_force_parallel_query_dop 1 set observer _force_parallel_query_dop variables false SYS obproxy_read_consistency 0 request target: 0 means leader; 1 means follower false SYS obproxy_read_only 1 read write type: 0 means ReadOnly, 1 means ReadWrite false SYS digest_sql_length 1024 0 means use default print sql len, otherwise is digest_sql_length false SYS enable_stat True if enabled, will collect stat info false SYS enable_trace True if enabled, log will print trace info false SYS enable_performance_mode False if enabled, for performance situation false SYS observer_sys_password1 password for observer sys user false SYS observer_sys_password password for observer sys user false SYS obproxy_sys_password 5d82f3ba321adee866da4999d5a33042bf5a4a54 password for obproxy sys user false SYS inspector_password password for inspector user false SYS citadel_agent_ip bkmi citadel agent ip false USER bt_mode beyond trust mode false USER bt_retry_times 3 beyond trust sdk retry times true USER bt_server_addr beyond trust server address or hostname false USER domain_name app domain name false USER session_pool_stat_log_interval 1m pool stat log interval, [0s, 1d] false USER session_pool_stat_log_ratio 9000 the num when reach will log false USER session_pool_default_prefill False session_pool_default_prefill false USER session_pool_default_blocking_timeout 500ms session_pool_default_blocking_timeout, [0ms, 2s] false USER session_pool_default_idle_timeout 1800s session_pool_default_idle_timeout, [0s, 1d] false USER session_pool_default_max_conn 20 the num of max conn , [0, 100000] false USER session_pool_default_min_conn 0 the num of min conn , [0, 100000] false USER use_local_session_prop False if enabled means use_local_session prop false USER session_pool_cont_delay_interval 5ms session_pool_cont_delay_interval, [0s, 1d] false USER max_pending_num 100000 the num of conn in pending list, [0, 100000] false USER create_conn_cont_num 10 the num of create conn cont, [0, 1000] false USER refresh_server_cont_num 5 the num of refresh server cont, [0,1000] false USER session_pool_retry_interval 1ms session_pool_retry_interval, [0s, 1d] false USER need_release_after_tx False if enabled means release server session after transaction complete false USER enable_no_sharding_skip_real_conn False if enabled no sharding will use saved password check to skip real conn false USER enable_session_pool_for_no_sharding False if enabled can use session pool for no sharding false USER enable_conn_precreate False if enabled means precreate conn for session pool false USER is_pool_mode False if enabled means useing session pool false USER auto_scan_all False if enabled, need scan all false USER cloud_instance_id ob sharding cloud instance id false USER mng_url ob sharding console url false USER runtime_env runtime env false USER instance_ip ip addr string false USER pod_namespace sidecar pod namespace false USER pod_name sidecar pod name false USER server_zone app server location false USER workspace_name app workspace name false USER env_tenant_name app tenant name false USER grpc_timeout 30m grpc client timeout, [1s, 1d] false USER enable_shard_authority False if enabled, check authority for sharding user false USER use_local_dbconfig False if enabled, start dbmesh with local dbconfig false USER dataplane_host dataplane address or hostname false USER sidecar_node_id node id for dbmesh false USER enable_sharding False if enabled means use logic db false USER enable_mysql_proxy_pool True if enabled, will long conn for sequence false USER sequence_prefetch_threshold 7000 when cost reach threshold will prefetch, [0, 10000] false USER enable_sequence_prefetch True if enabled, will prefetch sequence false USER sequence_fail_retry_count 2 the count sequence retry when fail, [0, 100] false USER sequence_entry_expire_time 1d sequence entry valid time, [0s, 1d] false USER enable_async_pull_location_cache True enable async pull location cache when is dirty false SYS check_tenant_locality_change True enable locality change trigger location cache dirty false SYS ldg_info_refresh_interval 20s config server info refresh task interval, [10s, 1d] false USER enable_ldg False enable proxy to support ldg false SYS enable_primary_zone True enable proxy route according to the tenant primary zone priority while proxy calculate route failed. false USER proxy_primary_zone_name primary zone name for proxy ldc route. If not empty, proxy only route to the zone false SYS enable_standby True enable standby or not false USER qos_stat_item_limit 3000 obproxy qos stat item false SYS qos_stat_item_max_idle_period 30m qos stat item in memory idle period. it will remove if timeout false USER qos_stat_clean_interval 1m clean qos stat interval, [1s, 1h], 0 means disable, if set a negative value, proxy treat it as 0 false USER enable_qos False if enabled, proxy will be able to qos false USER ssl_attributes store ssl config to control ssl behavior, works for new connection false SYS enable_server_ssl False if enabled, proxy will try best to connect server whith ssl false SYS enable_client_ssl False if enabled, proxy will try best to connect client with ssl false SYS skip_proxy_sys_private_check True skip_proxy_sys_private_check false USER skip_proxyro_check False used for proxro@sys, if set false, access denied false SYS enable_full_username True used for non-cloud user, if set true, username must have tenant and cluster false SYS enable_cloud_full_username False used for cloud user, if set false, treat all login user as username false SYS sql_table_cache_mem_limited 128MB max size of proxy sql table cache size. [1KB, 100G] false USER sql_table_cache_expire_relative_time 0 the unit is ms, 0 means do not expire, others will expire sql table cache base on relative time false SYS enable_index_route False enable index route or not false USER mysql_version 5.6.25 returned version for mysql mode, default value is 5.6.25. If set, proxy will send new version when user connect to proxy false USER enable_cpu_isolate False enable cpu isolate or not true USER proxy_route_policy follower_first proxy route policy false SYS qa_mode_mock_public_cloud_vid 1 mock public cloud vid false USER qa_mode_mock_public_cloud_slb_addr 127.0.0.1:33045 mock public cloud slb addr false USER location_expire_period_time 0d time for location expire period, values in [0s, 30d], 0 means no expire false SYS location_expire_period 0 just for test, not recommended, the unit is ms, only work if qa_mode is set, it means location cache which has been created for more than this value will be expired false SYS enable_qa_mode False just for test, not recommended, if enabled, proxy can forcibly expire all location cache false SYS partition_location_expire_relative_time 0 the unit is ms, 0 means do not expire, others will expire partition location base on relative time false SYS enable_causal_order_read True if enabled, proxy will choose server by priority and sync safe snapshot version if need false USER enable_extra_prometheus_metric False enable net and route prometheus merics or not false USER enable_prometheus True enable prometheus or not false USER prometheus_sync_interval 1s update sync metrics to prometheus exposer interval, [1s, 1h], 0 means disable, if set a negative value, proxy treat it as 0 false USER prometheus_cost_ms_unit False enable prometheus cost unit is ms, default is us true SYS prometheus_listen_port 2884 obproxy prometheus listen port true SYS enable_monitor_stat True enable monitor stat or not false USER monitor_stat_high_threshold 500ms tenant stat time high threshold false USER monitor_stat_middle_threshold 100ms tenant stat time middle threshold false USER monitor_stat_low_threshold 30ms tenant stat time low threshold false USER monitor_stat_dump_interval 1m dump monitor statistic in log interval, [1s, 1h], 0 means disable, if set a negative value, proxy treat it as 0 false USER monitor_cost_ms_unit False enable monitor cost unit is ms, default is us false SYS monitor_item_max_idle_period 30m monitor stat item in memory idle period. it will remove if timeout false USER monitor_item_limit 3000 obproxy monitor stat item/prometheus metric limit false SYS tenant_location_valid_time 1d tenant location valid time, [0s, 100d] false USER delay_update_entry_interval 5s delay update table entry or partition entry interval, [0s, 1d] false USER client_max_memory_size 8MB max dynamic alloc memory size of one client session false SYS max_connections 60000 max fd proxy could use false SYS local_vip_tenant_version 0 local vip tenant version false MEMORY current_local_config_version 0 local config version for current app false SYS proxy_idc_name idc3 idc name for proxy ldc route. If is empty or invalid, treat as do not use ldc. User session vars 'proxy_session_ldc' can cover it false SYS proxy_local_cmd 0 proxy local cmd type: 0->none(default), 1->exit, 2->restart, 3->offline, 4->online false MEMORY enable_async_log True if enabled, use async logging way, maybe lost some log when busy false USER xflush_log_level INFO specifies the current level of logging: DEBUG, TRACE, INFO, WARN, USER_ERR, ERROR false USER monitor_log_level INFO specifies the current level of logging: DEBUG, TRACE, INFO, WARN, USER_ERR, ERROR false USER syslog_level INFO specifies the current level of logging: DEBUG, TRACE, INFO, WARN, USER_ERR, ERROR false USER server_routing_mode oceanbase server routing mode: 1.oceanbase(default mode); 2.random; 3.mock; 4.mysql false SYS test_server_addr proxy will choose this addr(if not empty) as observer addr forcibly, format ip1:sql_port1;ip2:sql_port2 false SYS route_diagnosis_level 2 '0' disable route diagnosis, greater level will collect more dedicated info of route false SYS target_db_server proxy will choose to route target db server addr forcibly, format ip0:port0,ip1:port1,ip2:port2;ip3:port3;ip4:port4 false SYS internal_cmd_mem_limited 16M internal cmd response memory limited, [0, 64MB], 0 means unlimited false USER sqlaudit_mem_limited 0 sqlaudit memory limited, [0, 1GB] false USER limited_pl_update_threshold 10 in merging state, max partition location update task processing per second false SYS normal_pl_update_threshold 100 max partition location update task processing per second false SYS enable_cached_server True if enabled, use cached server session when no table entry false USER enable_pl_route True if enabled, pl will be accurate routing false USER enable_weak_reroute True if this and protocol_v2 enabled, proxy will reroute weak read request when server false USER enable_reroute False if this and protocol_v2 enabled, proxy will reroute when routing error false USER enable_transaction_internal_routing True if enabled, proxy will route the dml statement in a transaction to different servers false USER enable_ob_protocol_v2_with_client False if enabled, proxy will use oceanbase protocol 2.0 with client false USER enable_ob_protocol_v2 True if enabled, proxy will use oceanbase protocol 2.0 with server false USER enable_compression_protocol True if enabled, proxy will use compression protocol with server false USER enable_partition_table_route True if enabled, partition table will be accurate routing false USER enable_bad_route_reject False if enabled, bad route request will be rejected, e.g. first statement of transaction opened by BEGIN(or START TRANSACTION) without table name false USER detect_server_timeout 5s detect server sql timeout, [1s, 1h] false USER server_detect_refresh_interval 1s the interval to refresh server state for getting zone or server newest state, [10ms, 1h] false USER server_detect_fail_threshold 3 server detect try times false SYS server_detect_mode 1 0 means no detect, 1 means target server detect, 2 means all server detect false USER enable_congestion True enable congestion feature or not false USER min_congested_connect_timeout 100ms if client connect timeout after the time, proxy set target server alive congested, [1ms, 1h] false USER congestion_retry_interval 20s congestion retry interval, [1s, 1h] false USER congestion_fail_window 120s congestion failure window size, [1s, 1h] false USER min_keep_congestion_interval 20s minimum keep congestion interval, [1s, 1d] false USER congestion_failure_threshold 5 congestion failure threshold, [0, +∞], 0 means disable congestion false USER proxy_tenant_name default tenant name for cloud user false SYS rootservice_list a list of servers against which election candidate is checked for validation, format ip1:sql_port1;ip2:sql_port2 true SYS rootservice_cluster_name undefined default cluster name for rootservice_list true SYS enable_get_rslist_remote False enable direct get rootservice list from remote false SYS enable_metadb_used False use MetaDataBase when proxy run true SYS app_name ob_pro_obproxy_rw current application name which proxy works for, need defined, only modified when restart true SYS proxy_id 0 used to identify each obproxy, it can not be zero if proxy_service_mode is server true SYS proxy_service_mode client proxy deploy and service mode: 1.client(default); 2.server true SYS obproxy_config_server_url http://10.252.158.68:8080/services?User_ID=alibaba&UID=test&Action=GetObProxyConfig&ObproxyClusterName=ob_pro_obproxy_rw url of config info(rs list and so on) true SYS local_bound_ipv6_ip :: local bound ipv6 ip(any) true SYS ip_listen_mode 1 1 means ipv4 listen mode, 2 means ipv6 listen mode, 3 means ipv4 and ipv6 listen mode true SYS listen_port 2883 obproxy listen port true SYS local_bound_ip 0.0.0.0 local bound ip(any) true SYS ignore_local_config True ignore all local cached files, start proxy with remote json false USER slow_query_time_threshold 500ms slow query time threshold, [0s, 30d] false USER query_digest_time_threshold 100ms digest time threshold, [0s, 30d] false USER slow_proxy_process_time_threshold 2ms slow proxy process time threshold, [0s, 30d] false USER slow_transaction_time_threshold 1s slow transaction time threshold, [0s, 30d], if set a negative value, proxy treat it as 0 false USER enable_trace_stats False enable mysql trace stats false USER enable_cpu_topology False enable cpu topology, work threads bind to cpu true USER enable_strict_stat_time True enable strict statistic time, use gettimeofday or clock_gettime(CLOCK_REALTIME) false USER enable_report_session_stats False enable report client session statistic table false USER enable_mysqlsm_info True enable mysqlsm info, not used in proxy now false USER enable_sync_all_stats True enable proxy sync all stats false USER enable_trans_detail_stats True enable mysql transaction detail stats false USER enable_flow_control True whether flow control is enabled in a mysql tunnel, applied instantly in new created tunnels after updated false USER flow_event_queue_threshold 5 event queue threshold for flow control, [0, 20], if set a negative value, proxy treat it as 5 false USER flow_consumer_reenable_threshold 256 consumer reenable threshold for flow control, [0, 131072], if set a negative value, proxy treat it as 256 false USER flow_low_water_mark 64K flow low water mark for flow control, [0, 16MB], if set a negative value, proxy treat it as 64K false USER flow_high_water_mark 64K flow high water mark for flow control, [0, 16MB], if set a negative value, proxy treat it as 64K false USER request_buffer_length 4KB the max length of request buffer we will alloc for each reqeust false USER tunnel_request_size_threshold 8KB use tunnel to transfer request, [4KB, 16MB], if request bigger than the threshold, 0 disable false USER default_buffer_water_mark 32KB default buffer water mark, [4B, 64KB] false USER max_log_file_size 256MB max size of log file, [1MB, 1G] false USER enable_strict_kernel_release False If is true, proxy only support 5u/6u/7u redhat. Otherwise no care kernel release, and proxy maybe unstable true SYS automatic_match_work_thread True ignore work_thread_num configuration item, use the count of cpu for current proxy work thread num true USER shard_scan_thread_num 0 proxy shard scan thread num, if 0, use (real work thread num/2), [0, 128] true USER grpc_client_num 9 proxy grpc client num, [9, 16] true USER grpc_thread_num 8 proxy grpc thread num, [8, 16] true USER block_thread_num 1 proxy block thread num, [1, 4] true USER task_thread_num 2 proxy task thread num, [1, 4] true USER work_thread_num 128 proxy work thread num or max work thread num when automatic match, [1, 128] true USER routing_cache_mem_limited 128MB max size of all proxy routing cache size, like table cache, location cache, etc. [1KB, 100G] false USER stack_size 1MB stack size of one thread, [1MB, 10MB] true USER proxy_mem_limited 3G proxy memory limited, [100MB, 100G] false USER client_tcp_user_timeout 0 client tcp user timeout, unit is s, 0 means no user timeout false USER client_tcp_keepcnt 2 client tcp keepalive probe count, 0 means use default value by kernel false USER client_tcp_keepintvl 5 client tcp keepalive interval time, unit is second, 0 means use default value by kernel false USER client_tcp_keepidle 5 client tcp keepalive idle time, unit is second, 0 means use default value by kernel false USER client_sock_option_flag_out 3 client sock param, option flag out, bit 1: NO_DELAY, bit 2: KEEP_ALIVE false USER server_tcp_user_timeout 0 tcp user timeout, unit is s, 0 means no user timeout false USER server_tcp_keepcnt 2 tcp keepalive probe count, 0 means use default value by kernel false USER server_tcp_keepintvl 5 tcp keepalive interval time, unit is second, 0 means use default value by kernel false USER server_tcp_keepidle 5 tcp keepalive idle time, unit is second, 0 means use default value by kernel false USER server_tcp_init_cwnd 0 the initial tcp congestion window, [0, 64] false USER sock_packet_tos_out 0 sock param, packet tos out, [0, 1] false USER sock_packet_mark_out 0 sock param, packet mark out, [0, 1] false USER sock_option_flag_out 3 sock param, option flag out, bit 1: NO_DELAY, bit 2: KEEP_ALIVE, bit 3: LINGER_ON false USER sock_send_buffer_size_out 0 sock param, send buffer size, [0, 8MB], if set a negative value, proxy treat it as 0 false USER sock_recv_buffer_size_out 0 sock param, recv buffer size, [0, 8MB], if set a negative value, proxy treat it as 0 false USER default_inactivity_timeout 180000s default inactivity timeout, [1s, 30d] false USER net_config_poll_timeout 1ms not used, just for compatible false USER net_accept_threads 2 net accept threads num, [0, 8] true USER frequent_accept True frequent accept true USER connect_observer_max_retries 3 max retries to do connect false USER enable_abort_conn_info False if enabled, proxy will send error packet when abort connection false USER enable_client_ip_checkout True if enabled, proxy send client ip when login false USER enable_proxy_scramble True if enable proxy scramble, proxy will send client its variable scramble num, not support old observer false SYS enable_cluster_checkout True if enable cluster checkout, proxy will send cluster name when login and server will check it false USER observer_query_timeout_delta 20s the delta value for @@ob_query_timeout, to cover net round trip time(proxy<->server) and task schedule time(server), [1s, 30s] false USER client_max_connections 8192 client max connections for one obproxy, [0, +∞] false USER enable_client_connection_lru_disconnect False if client connections reach throttle, true is that new connection will be accepted, and eliminate lru client connection, false is that new connection will disconnect, and err packet will be returned false SYS username_separator :;-;. username separator false SYS short_async_task_timeout 5s short async task timeout, [1s, 1h] false USER long_async_task_timeout 60s long async task timeout, [1s, 1h] false USER need_convert_vip_to_tname False convert vip to tenant name, which is useful in cloud false SYS enable_syslog_file_compress False Whether to enable archive log compression false SYS max_syslog_file_count 0 Maximum number of archive files to keep, 0 means ignore such limit false USER max_syslog_file_time 7d Maximum retention time of archive logs, 0 means ignore such limit false USER log_dir_size_threshold 10G max usable space size of log dir, used to decide whether should clean up log file, [256MB, 1T] false USER log_cleanup_interval 1m log file clean up task schedule interval, set 1 day or longer, [5s, 30d] false USER log_file_percentage 80 max percentage of avail size occupied by proxy log file, [0, 90], 0 means ignore such limit false USER delay_exit_time 100ms delay exit time, [100ms,500ms] false USER hot_upgrade_exit_timeout 30000000 graceful exit timeout, also use as INACTIVE timeout when rolling upgrade, unit is us, default 30s, -1 means no timeout, 0 means quit now, > 0 means wait time false USER hot_upgrade_rollback_timeout 24h default hot upgrade rollback timeout, proxy will do rollback if receive no rollback command in such long time, [1s, 30d] false USER hot_upgrade_failure_retries 5 default hot upgrade failure retries, proxy will stop handle hot_upgrade command after such retries, [1, 20] false USER fetch_proxy_bin_timeout 120s default hot upgrade fetch binary timeout, proxy will stop fetching after such long time, [1s, 1200s] false USER fetch_proxy_bin_random_time 300s max random waiting time of fetching proxy bin in hot upgrade, [1s, 1h] false USER cluster_expire_time 1d cluster resource expire time, 0 means never expire,cluster will be deleted if it has not been accessed for more than the time,[0, ] false USER cluster_count_high_water_mark 256 if cluster count is greater than this water mark, cluser will be kicked out by LRU false USER stat_dump_interval 6000s dump statistic in log interval, [0s, 1d], 0 means disable, if set a negative value, proxy treat it as 0 false USER stat_table_sync_interval 60s update sync statistic to ob_all_proxy_stat table interval, [0s, 1d], 0 means disable, if set a negative value, proxy treat it as 0 false USER idc_list_refresh_interval 2h the interval to refresh idc list for getting newest region-idc, [10s, 1d] false USER config_server_refresh_interval 60s config server info refresh task interval, [10s, 1d] false USER metadb_batch_interval 10ms pull config from metadb, replace interval time false USER metadb_batch_size 5 pull config from metadb, replace batch size to proxy_config false USER metadb_server_state_refresh_interval 60s the interval to refresh metadb server state for getting zone or server newest state, [10ms, 1h] false USER server_state_refresh_interval 20s the interval to refresh server state for getting zone or server newest state, [10ms, 1h] false USER cache_cleaner_clean_interval 20s the interval for cache cleaner to clean cache, [1s, 1d] false USER proxy_hot_upgrade_check_interval 5s proxy info check task interval, [1s, 1h] false USER proxy_info_check_interval 60s proxy info check task interval, [1s, 1h] false USER refresh_config False when table processor do check work, update all proxy config if refresh_config is true false MEMORY refresh_idc_list False when refresh config server, update all idc list if refresh_idc_list is true false MEMORY refresh_rslist False when refresh config server, update all rslist if refresh_rslist is true false MEMORY refresh_json_config False force update json info if refresh_json_config is true false MEMORY enable_xa_route True whether to enable hold XA_START and route to the partition that the next dml will visit false USER json_config_version 9b6a346dcffe8dfcad151da5fd5c78de json config info version false virtual json_config_bin_url http://10.252.158.68:8080/client?Action=GetObProxy&User_ID=alibaba&UID=test bin url, used for hot upgrade false virtual json_config_meta_table_db *** meta table db name true virtual json_config_meta_table_user *** meta table username, format'user@tenant#cluster' true virtual json_config_meta_table_password *** meta table password true virtual json_config_real_meta_cluster meta table real cluster true virtual json_config_cluster_count 1 ob cluster count, meta db cluster not included false virtual json_config_modified_time 2024-08-22 16:08:38.331464 json config modified time false virtual