The following table describes ODP parameters that take effect immediately after modification.
| Parameter | Default value | Value range | Description |
|---|---|---|---|
mysqlclient_query_timeout |
120s | [1s, +∞) | Specifies the query timeout when ODP uses a MySQL client to access OceanBase Database. |
mysqlclient_net_wait_timeout |
10s | [-1, +∞) | Specifies the connection timeout when ODP uses a MySQL client to access OceanBase Database. It includes read_timeout and write_time_out. You can set the value to -1 to make connections never time out. |
mysqlclient_connection_refresh_interval |
200ms | [200ms, +∞) | Specifies the time interval to refresh the MySQL connection pool. The actual interval is the configured time × 50. |
mysqlclient_connection_pool_warn_time |
1s | [0s, +∞) | Specifies the time period that triggers a connection pool alert. If a connection has been used for a long time, an alert is registered in the log. |
mysqlclient_conn_limit_per_observer |
10 | [1, +∞) | Specifies the maximum number of SQL server connections that a connection pool can handle. |
proxy_info_check_interval |
60s | [1s, 1h] | Specifies the time interval at which ODP checks for configuration updates. |
| cache_cleaner_clean_interval | 20s | [1s, 1d] | Specifies the time interval at which the cache cleaner cleans the cache. |
schema_refresh_interval |
20s | [1s, 1h] | Specifies the time interval at which the schema is refreshed. |
server_state_refresh_interval |
20s | (0s, 1h] | Specifies the time interval at which the process status of OceanBase Database is refreshed. |
| metadb_server_state_refresh_interval | 60s | [10ms, 1h] | Specifies the time interval at which the MetaDB is refreshed. |
config_server_refresh_interval |
20s | [60s, 1d] | Specifies the time interval at which the config server is refreshed. |
| idc_list_refresh_interval | 2h | [10s, 1d] | Specifies the time interval at which the IDC information is refreshed. |
stat_sync_interval |
60s | [0s, 1d] | Specifies the time interval at which the global statistics synchronization task is run. You can set the value to 0 to disable the task. |
stat_dump_interval |
6000s | [0s, 1d] | Specifies the time interval at which the global statistics dumping task is executed. |
stat_table_sync_interval |
60s | [0s, 1d] | Specifies the time interval at which the global statistics on internal tables are updated. |
| cluster_count_high_water_mark | 256MB | [2, 102400] | Specifies the maximum size of resources that can be created in the cluster. |
| cluster_expire_time | 1d | [0, +∞) | Specifies the idle timeout of cluster metadata. |
fetch_proxy_bin_random_time |
300s | [1s, 1h] | Specifies the maximum waiting time before ODP can fetch new binary files in a hot upgrade. |
hot_upgrade_rollback_timeout |
24h | [1s, 30d] | Specifies the default waiting time to execute a rollback after a hot upgrade. |
hot_upgrade_failure_retries |
5 | [1, 20] | Specifies the number of retries after a hot upgrade failed. |
fetch_proxy_bin_timeout |
120s | [1s, 1200s] | Specifies the timeout when fetching new binary files in a hot upgrade. |
hot_upgrade_graceful_exit_timeout |
120s | [0s, 30d] | Specifies the waiting time before ODP automatically exits. |
| delay_exit_time | 100ms | [100ms, 500ms] | Specifies the amount of time by which an ODP exit is delayed. |
log_file_percentage |
80 | [0, 100] | Specifies the percentage threshold of ODP logs. Log files are cleaned up when the threshold is exceeded. |
log_cleanup_interval |
3600s | [5s, 30d] | Specifies the time interval at which the cleanup task is executed. |
log_dir_size_threshold |
64GB | [256MB, 1T] | Specifies the size threshold of ODP logs. Log files are cleaned up when the threshold is exceeded. |
max_log_file_size |
256M | [1MB, 1G] | Specifies the maximum size of a log file. |
long_async_task_timeout |
60s | [1s, 1h] | Specifies the timeout of long asynchronous tasks. |
short_async_task_timeout |
5s | [1s, 1h] | Specifies the timeout of short asynchronous tasks. |
client_max_connections |
8192 | [0, 65535] | Specifies the maximum number of client connections supported by ODP. |
observer_query_timeout_delta |
20s | [1s, 30s] | Specifies the ob_query_timeout incremental value for the OceanBase Database process. |
| enable_cluster_checkout | true | / | Specifies whether to verify the cluster name. |
| enable_client_ip_checkout | true | / | Specifies whether to verify the cluster IP address. |
connect_observer_max_retries |
3 | [2, 5] | Specifies the maximum number of retries when ODP connects to the OceanBase Database process. |
net_config_poll_timeout |
10ms | [0, +∞) | Specifies the timeout for network events. |
default_inactivity_timeout |
180000s | [1s, 30d] | Specifies the default timeout for the TCP connection. |
sock_send_buffer_size_out |
0 | [0, 8MB] | Specifies the size of a socket send buffer. |
sock_recv_buffer_size_out |
0 | [0, 8MB] | Specifies the size of a socket receive buffer. |
sock_option_flag_out |
0 | [0, 1] | Specifies the socket flag. |
sock_packet_mark_out |
0 | [0, 1] | Specifies the socket mark. |
sock_packet_tos_out |
0 | [0, 1] | Specifies the socket TOS. |
server_tcp_init_cwnd |
0 | [0, 64] | Specifies the initial TCP congestion window when creating a new server connection. |
server_tcp_keepidle |
0 | [0, 7200] | Specifies the TCP keepalive time for the connection between ODP and the OceanBase Database process. |
server_tcp_keepintvl |
0 | [0, 75] | Specifies the TCP keepalive interval for the connection between ODP and the OceanBase Database process. |
server_tcp_keepcnt |
0 | [0, 9] | Specifies the TCP keepalive probe count for the connection between ODP and the OceanBase Database process. |
server_tcp_user_timeout |
5 | [0, 20] | Specifies the TCP user timeout for the connection between ODP and the OceanBase Database process. |
client_sock_option_flag_out |
0 | [0, +∞) | Specifies the socket flag for the connection between ODP and the client. |
client_tcp_keepidle |
0 | [0, 7200] | Specifies the TCP keepalive time for the connection between ODP and the client. |
client_tcp_keepintvl |
0 | [0, 75] | Specifies the TCP keepalive interval for the connection between ODP and the client. |
client_tcp_keepcnt |
0 | [0, 9] | Specifies the TCP keepalive probe count for the connection between ODP and the client. |
client_tcp_user_timeout |
5 | [0, 20] | Specifies the TCP user timeout for the connection between ODP and the client. |
proxy_mem_limited |
2G | [100MB, 100GB] | Specifies the maximum memory usage during the ODP runtime. ODP automatically exits when the limit is exceeded. |
routing_cache_mem_limited |
128M | [1KB, 100G] | Specifies the maximum memory usage for the routing cache of PL/SQL queries. |
enable_flow_control |
false | / | Specifies whether to enable throttling in the tunnel. |
flow_high_water_mark |
0 | [0, 16MB] | Specifies the upper limit of tunnel throttling. |
flow_low_water_mark |
0 | [0, 16MB] | Specifies the lower limit of tunnel throttling. |
flow_consumer_reenable_threshold |
0 | [0, 131072] | Specifies the tunnel throttling threshold on the number of data forwards. |
flow_event_queue_threshold |
0 | [0, 20] | Specifies the tunnel throttling threshold on the number of tasks in the worker thread queue. |
tunnel_request_size_threshold |
84KB | [4KB, 16MB] | Specifies the threshold of requests received by ODP. When the size of client requests exceeds this threshold, ODP uses the tunnel to forward them and does not parse them. |
request_buffer_length |
4KB | [1KB, 16MB] | Specifies the size of the request buffer. |
default_buffer_size |
8KB | [1KB, 64KB] | Specifies the default size of the request and response blocks. |
default_buffer_water_mark |
32KB | [4B, 64KB] | Specifies the default buffer high-water mark. |
enable_trans_detail_stats |
true | / | Specifies whether to enable statistics collection of transaction status. |
| enable_sync_all_stats | true | / | Specifies whether to synchronize all statistics. |
enable_mysqlsm_info |
true | / | |
enable_report_session_stats |
true | / | Specifies whether to enable the reporting of session-wide statistics to the internal table. |
enable_strict_stat_time |
true | / | Specifies whether to enable the strict statistic time. |
enable_trace_stats |
false | / | Specifies whether to enable MySQL tracing. |
slow_transaction_time_threshold |
5s | [0s, 30d] | Specifies the transaction execution time threshold. When the execution time exceeds the threshold, an alert is registered in the log. |
slow_proxy_process_time_threshold |
2ms | [0s, 30d] | Specifies the execution time threshold of the ODP pre-execution request. When the execution time exceeds the threshold, an alert is registered in the log. |
query_digest_time_threshold |
100ms | [0s, 30d] | Specifies the execution time threshold of requests. When the execution time exceeds the threshold, a log record is registered to the obproxy_digest.log file. |
slow_query_time_threshold |
500ms | [0s, 30d] | Specifies the execution time threshold of slow requests. When the execution time exceeds the threshold, a log record is registered to the obproxy_slow.log file. |
ignore_local_config |
true | / | Specifies whether to ignore the local configuration file. |
congestion_failure_threshold |
5 | [0, +∞) | Specifies the threshold to blacklist a faulty OBServer. If the number of OBServer errors within the period of congestion_fail_window exceeds the threshold, ODP adds the server to the blacklist. |
min_keep_congestion_interval |
20s | [1s, 1d] | Specifies the minimum waiting time to unblacklist an OBServer. |
congestion_fail_window |
120s | [1s, 1h] | Specifies the statistical period for collecting OBServer errors. This parameter is used in conjunction with congestion_failure_threshold. |
congestion_retry_interval |
20s | [1s, 1h] | Specifies the time interval to retry the blacklisted OBServers that are alive but disabled. |
min_congested_connect_timeout |
100ms | [1ms, 1h] | Specifies the minimum timeout for retrying the blacklisted OBServers that are alive but disabled. |
enable_congestion |
true | / | Specifies whether to enable the blacklist. |
enable_bad_route_reject |
false | / | Specifies whether to reject requests that cannot be routed when the blacklist is enabled. |
enable_partition_table_route |
true | / | Specifies whether to enable routing for partitioned tables. |
enable_reroute |
false | / | Specifies whether to enable rerouting. |
enable_pl_route |
false | / | Specifies whether to enable Procedural Language (PL) routing. |
syslog_level |
INFO | DEBUG, TRACE, INFO, WARN, USER_ERR, ERROR | Specifies the log level. |
monitor_log_level |
INFO | DEBUG, TRACE, INFO, WARN, USER_ERR, ERROR | Specifies the monitoring log level. |
xflush_log_level |
INFO | DEBUG, TRACE, INFO, WARN, USER_ERR, ERROR | Specifies the xflush log level. |
enable_async_log |
true | / | Specifies whether to enable or disable the asynchronous logs. |