- name: home_path require: true type: STRING min_value: NULL max_value: NULL need_redeploy: true description_en: the directory for the work data file description_local: OceanBase工作目录 - name: cluster_id require: true type: INT default: 1 min_value: 1 max_value: 4294901759 modify_limit: modify need_restart: true description_en: ID of the cluster description_local: 本OceanBase集群ID - name: data_dir type: STRING min_value: NULL max_value: NULL need_redeploy: true description_en: the directory for the data file description_local: 存储sstable等数据的目录 - name: redo_dir type: STRING min_value: NULL max_value: NULL need_redeploy: true description_en: the directory for the redo file description_local: 存储clog, iclog, slog数据的目录 - name: clog_dir type: STRING min_value: NULL max_value: NULL need_redeploy: true description_en: the directory for the clog file description_local: 存储clog数据的目录, clog 应该与 ilog 同盘 - name: slog_dir type: STRING min_value: NULL max_value: NULL need_redeploy: true description_en: the directory for the slog file description_local: 存储slog数据的目录. 4.0版本开始不支持配置该项 - name: ilog_dir type: STRING min_value: NULL max_value: NULL need_redeploy: true description_en: the directory for the ilog file description_local: 存储ilog数据的目录 - name: devname type: STRING min_value: NULL max_value: NULL need_restart: true description_en: name of network adapter description_local: 服务进程绑定的网卡设备名 - name: rpc_port require: true type: INT default: 2882 min_value: 1025 max_value: 65535 modify_limit: modify need_restart: true description_en: the port number for RPC protocol. description_local: 集群内部通信的端口号 - name: mysql_port require: true type: INT default: 2881 min_value: 1025 max_value: 65535 modify_limit: modify need_restart: true description_en: port number for mysql connection description_local: SQL服务协议端口号 - name: zone require: true type: STRING default: zone1 min_value: NULL max_value: NULL section: OBSERVER need_redeploy: true description_en: specifies the zone name description_local: 节点所在的zone的名字。 - name: max_px_worker_count require: false type: INT default: 64 min_value: 0 max_value: 65535 section: TENANT need_restart: false description_en: maximum parallel execution worker count can be used for all parallel requests. description_local: SQL并行查询引擎使用的最大线程数 - name: sys_cpu_limit_trigger require: false type: INT default: 80 min_value: 50 max_value: NULL section: OBSERVER need_restart: false description_en: when the cpu usage percentage exceed the trigger, will limit the sys cpu usage description_local: 当CPU利用率超过该阈值的时候，将暂停系统后台任务的执行 - name: memory_limit_percentage require: false type: INT default: 80 min_value: 10 max_value: 90 modify_limit: decrease section: OBSERVER need_restart: false description_en: memory limit percentage of the total physical memory description_local: 系统总可用内存大小占总内存大小的百分比 - name: sys_bkgd_migration_retry_num require: false type: INT default: 3 min_value: 3 max_value: 100 section: OBSERVER need_restart: false description_en: retry num limit during migration. description_local: 副本迁移失败时最多重试次数 - name: tableapi_transport_compress_func require: false type: STRING default: none min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: compressor used for tableAPI query result. description_local: tableAPI查询结果传输使用的压缩算法 - name: disk_io_thread_count require: false type: INT default: 8 min_value: 2 max_value: 32 section: OBSERVER need_restart: false description_en: The number of io threads on each disk. description_local: 磁盘IO线程数。必须为偶数。 - name: location_cache_refresh_min_interval require: false type: TIME default: 100ms min_value: 0s max_value: NULL section: LOCATION_CACHE need_restart: false description_en: the time interval in which no request for location cache renewal will be executed. description_local: 位置缓存刷新请求的最小间隔，防止产生过多刷新请求造成系统压力过大 - name: trace_log_slow_query_watermark type: TIME default: 1s min_value: 1ms max_value: NULL section: OBSERVER need_restart: false description_en: the threshold of execution time (in milliseconds) of a query beyond which it is considered to be a slow query. description_local: 执行时间超过该阈值的查询会被认为是慢查询，慢查询的追踪日志会被打印到系统日志中 - name: max_string_print_length require: false type: INT default: 500 min_value: 0 max_value: NULL section: OBSERVER need_restart: false description_en: truncate very long string when printing to log file description_local: 打印系统日志时，单行日志最大长度 - name: row_compaction_update_limit require: false type: INT default: 6 min_value: 1 max_value: 6400 section: TRANS need_restart: false description_en: maximum update count before trigger row compaction description_local: 触发内存中行内数据合并的修改次数 - name: enable_rereplication require: false type: BOOL default: true min_value: NULL max_value: NULL section: LOAD_BALANCE need_restart: false description_en: specifies whether the partition auto-replication is turned on. description_local: 自动补副本开关 - name: rootservice_async_task_thread_count require: false type: INT default: 4 min_value: 1 max_value: 10 section: ROOT_SERVICE need_restart: false description_en: maximum of threads allowed for executing asynchronous task at rootserver. description_local: RootService内部异步任务使用的线程池大小 - name: major_compact_trigger require: false type: INT default: 5 min_value: 0 max_value: 65535 section: TENANT need_restart: false description_en: major_compact_trigger alias to minor_freeze_times description_local: 多少次小合并触发一次全局合并。值为0时，表示关闭小合并 - name: default_compress require: false type: STRING default: archive min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: default compress function name for create new table description_local: Oracle模式下，建表时使用的默认压缩策略 - name: ssl_client_authentication require: false type: BOOL default: false min_value: NULL max_value: NULL section: OBSERVER need_restart: true description_en: enable server supports SSL connection, takes effect only after server restart with all ca/cert/key file. description_local: 是否开启SSL连接功能 - name: datafile_size require: false type: CAPACITY default: 0 min_value: 0M max_value: NULL modify_limit: decrease section: SSTABLE need_restart: false description_en: size of the data file. description_local: 数据文件大小。一般不要设置。 - name: merge_stat_sampling_ratio require: false type: INT default: 100 min_value: 0 max_value: 100 section: OBSERVER need_restart: false description_en: column stats sampling ratio daily merge. description_local: 合并时候数据列统计信息的采样率 - name: cache_wash_threshold require: false type: CAPACITY default: 4GB min_value: 0B max_value: NULL section: OBSERVER need_restart: false description_en: size of remaining memory at which cache eviction will be triggered. description_local: 触发缓存清理的容量阈值 - name: user_iort_up_percentage require: false type: INT default: 100 min_value: 0 max_value: NULL section: OBSERVER need_restart: false description_en: variable to control sys io, the percentage of use io rt can raise description_local: 用户磁盘IO时延超过该阈值后，系统后台IO任务将被限流 - name: high_priority_net_thread_count require: false type: INT default: 0 min_value: 0 max_value: 100 section: OBSERVER need_restart: true description_en: the number of rpc I/O threads for high priority messages, 0 means set off description_local: 高优先级网络线程数，值0表示关闭 - name: max_kept_major_version_number require: false type: INT default: 2 min_value: 1 max_value: 16 section: DAILY_MERGE need_restart: false description_en: the maximum number of kept major versions description_local: 数据保留多少个冻结版本 - name: enable_sys_unit_standalone require: false type: BOOL default: false min_value: NULL max_value: NULL section: LOAD_BALANCE need_restart: false description_en: specifies whether sys unit standalone deployment is turned on. description_local: 系统租户UNIT是否独占节点 - name: freeze_trigger_percentage require: false type: INT default: 50 min_value: 1 max_value: 99 section: TENANT need_restart: false description_en: the threshold of the size of the mem store when freeze will be triggered. description_local: 触发全局冻结的租户使用内存阈值。另见enable_global_freeze_trigger。 - name: enable_major_freeze require: false type: BOOL default: true min_value: NULL max_value: NULL section: ROOT_SERVICE need_restart: false description_en: specifies whether major_freeze function is turned on. description_local: 自动全局冻结开关 - name: balancer_tolerance_percentage require: false type: INT default: 10 min_value: 1 max_value: 99 section: LOAD_BALANCE need_restart: false description_en: specifies the tolerance (in percentage) of the unbalance of the disk space utilization among all units. description_local: 租户内多个UNIT间磁盘不均衡程度的宽容度，在均值+-宽容度范围之内的不均衡不会触发执行均衡动作 - name: server_cpu_quota_min require: false type: DOUBLE default: 2.5 min_value: 0 max_value: 16 section: TENANT need_restart: true description_en: the number of minimal vCPUs allocated to the server tenant(a special internal tenant that exists on every observer) description_local: 系统可以使用的最小CPU配额，将会预留 - name: memory_reserved require: false type: CAPACITY default: 500M min_value: 10M max_value: NULL section: SSTABLE need_restart: false description_en: the size of the system memory reserved for emergency internal use. description_local: 系统预留内存大小 - name: server_cpu_quota_max require: false type: DOUBLE default: 5 min_value: 0 max_value: 16 section: TENANT need_restart: true description_en: the number of maximal vCPUs allocated to the server tenant description_local: 系统可以使用的最大CPU配额 - name: rootservice_ready_check_interval require: false type: TIME default: 3s min_value: 100000us max_value: 1m section: ROOT_SERVICE need_restart: false description_en: the interval between the schedule of the task that checks on the status of the ZONE during restarting. description_local: RootService启动后等待和检查集群状态的时间间隔 - name: debug_sync_timeout require: false type: TIME default: 0 min_value: 0 max_value: NULL section: OBSERVER need_restart: false description_en: Enable the debug sync facility and optionally specify a default wait timeout in micro seconds. A zero value keeps the facility disabled description_local: 打开debug sync调试开关，并设置其超时时间；值为0时，则关闭。 - name: syslog_level require: false type: STRING default: INFO min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies the current level of logging. description_local: 系统日志级别 - name: resource_hard_limit require: false type: INT default: 100 min_value: 1 max_value: 10000 section: LOAD_BALANCE need_restart: false description_en: Used along with resource_soft_limit in unit allocation. If server utilization is less than resource_soft_limit, a policy of best fit will be used for unit allocation; otherwise, a least load policy will be employed. Ultimately,system utilization should not be large than resource_hard_limit. description_local: CPU和内存等资源进行分配的时候，资源总量是实际数量乘以该百分比的值 - name: leak_mod_to_check require: false type: STRING default: NONE min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: the name of the module under memory leak checks description_local: 内存泄露检查，用于内部调试目的 - name: balancer_task_timeout require: false type: TIME default: 20m min_value: 1s max_value: NULL section: LOAD_BALANCE need_restart: false description_en: the time to execute the load-balancing task before it is terminated. description_local: 负载均衡等后台任务的超时时间 - name: enable_upgrade_mode require: false type: BOOL default: false min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether upgrade mode is turned on. If turned on, daily merger and balancer will be disabled. description_local: 升级模式开关。在升级模式中，会暂停部分系统后台功能。 - name: multiblock_read_size require: false type: CAPACITY default: 128K min_value: 0K max_value: 2M section: SSTABLE need_restart: false description_en: multiple block batch read size in one read io request. description_local: 读取数据时IO聚合大小 - name: migration_disable_time require: false type: TIME default: 3600s min_value: 1s max_value: NULL section: ROOT_SERVICE need_restart: false description_en: the duration in which the observer stays in the block_migrate_in status, which means no partition is allowed to migrate into the server. description_local: 因磁盘满等原因导致某个节点数据迁入失败时，暂停迁入时长 - name: tablet_size require: false type: CAPACITY default: 128M min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: default tablet size, has to be a multiple of 2M description_local: 分区内部并行处理（合并、查询等）时每个分片的大小 - name: dead_socket_detection_timeout require: false type: TIME default: 10s min_value: 0s max_value: 2h section: OBSERVER need_restart: false description_en: specify a tcp_user_timeout for RFC5482. A zero value makes the option disabled description_local: 失效socket检测超时时间 - name: server_check_interval require: false type: TIME default: 30s min_value: 1s max_value: NULL section: ROOT_SERVICE need_restart: false description_en: the time interval between schedules of a task that examines the __all_server table. description_local: server表一致性检查的时间间隔 - name: lease_time require: false type: TIME default: 10s min_value: 1s max_value: 5m section: ROOT_SERVICE need_restart: false description_en: Lease for current heartbeat. If the root server does not received any heartbeat from an observer in lease_time seconds, that observer is considered to be offline. description_local: RootService与其他服务节点之间的租约时长。一般请勿修改。 - name: rootservice_async_task_queue_size require: false type: INT default: 16384 min_value: 8 max_value: 131072 section: ROOT_SERVICE need_restart: false description_en: the size of the queue for all asynchronous tasks at rootserver. description_local: RootService内部异步任务队列的容量 - name: location_refresh_thread_count require: false type: INT default: 4 min_value: 2 max_value: 64 section: LOCATION_CACHE need_restart: false description_en: the number of threads that fetch the partition location information from the root service. description_local: 用于位置缓存刷新的线程数 - name: minor_compact_trigger require: false type: INT default: 2 min_value: 0 max_value: 16 section: TENANT need_restart: false description_en: minor_compact_trigger description_local: 触发小合并的迷你合并次数 - name: major_freeze_duty_time type: MOMENT default: Disable min_value: 00:00 max_value: 23:59 section: DAILY_MERGE need_restart: false description_en: the start time of system daily merge procedure. description_local: 每日定时冻结和合并的触发时刻 - name: ignore_replay_checksum_error require: false type: BOOL default: false min_value: NULL max_value: NULL section: TRANS need_restart: false description_en: specifies whether error raised from the memtable replay checksum validation can be ignored. description_local: 是否忽略回放事务日志时发生的校验和错误 - name: user_block_cache_priority require: false type: INT default: 1 min_value: 1 max_value: NULL section: CACHE need_restart: false description_en: user block cache priority description_local: 数据块缓存在缓存系统中的优先级 - name: syslog_io_bandwidth_limit require: false type: CAPACITY default: 30MB min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: Syslog IO bandwidth limitation, exceeding syslog would be truncated. Use 0 to disable ERROR log. description_local: 系统日志所能占用的磁盘IO带宽上限，超过带宽的系统日志将被丢弃 - name: workers_per_cpu_quota require: false type: INT default: 10 min_value: 2 max_value: 20 section: TENANT need_restart: false description_en: the ratio(integer) between the number of system allocated workers vs the maximum number of threads that can be scheduled concurrently. description_local: 每个CPU配额分配多少个工作线程 - name: enable_record_trace_id require: false type: BOOL default: true min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether record app trace id is turned on. description_local: 是否记录应用端设置的追踪ID - name: config_additional_dir require: false type: STRING_LIST default: etc2;etc3 min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: additional directories of configure file description_local: 本地存储配置文件的多个目录，为了冗余存储多份配置文件 - name: enable_syslog_recycle require: false type: BOOL default: false min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether log file recycling is turned on description_local: 是否自动回收系统日志 - name: px_task_size require: false type: CAPACITY default: 2M min_value: 2M max_value: NULL section: OBSERVER need_restart: false description_en: min task access size of px task description_local: SQL并行查询引擎每个任务处理的数据量大小 - name: replica_safe_remove_time require: false type: TIME default: 2h min_value: 1m max_value: NULL section: ROOT_SERVICE need_restart: false description_en: the time interval that replica not existed has not been modified beyond which a replica is considered can be safely removed description_local: 已删除副本可以被清理的安全保留时间 - name: builtin_db_data_verify_cycle require: false type: INT default: 20 min_value: 0 max_value: 360 section: OBSERVER need_restart: false description_en: check cycle of db data. description_local: 数据坏块自检周期，单位为天。值0表示不检查。 - name: system_cpu_quota require: false type: DOUBLE default: 10 min_value: 0 max_value: 16 section: TENANT need_restart: false description_en: the number of vCPUs allocated to the server tenant description_local: 系统后台任务可使用CPU配额 - name: enable_sys_table_ddl require: false type: BOOL default: false min_value: NULL max_value: NULL section: ROOT_SERVICE need_restart: false description_en: specifies whether a system table is allowed be to created manually. description_local: 是否允许新建和修改系统表。主要在系统升级过程中使用。 - name: merge_thread_count require: false type: INT default: 0 min_value: 0 max_value: 256 section: OBSERVER need_restart: false description_en: worker thread num for compaction description_local: 用于合并的线程数 - name: net_thread_count require: false type: INT default: 0 min_value: 0 max_value: 128 section: OBSERVER need_restart: true description_en: the number of rpc/mysql I/O threads for Libeasy. description_local: 网络IO线程数 - name: max_stale_time_for_weak_consistency require: false type: TIME default: 5s min_value: 5s max_value: NULL section: OBSERVER need_restart: false description_en: the max data stale time that observer can provide service when its parent is invalid. description_local: 弱一致性读允许读到多旧的数据 - name: backup_log_archive_option require: false type: STRING default: OPTIONAL min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: backup log archive option, support MANDATORY/OPTIONAL, COMPRESSION description_local: 日志备份的参数 - name: backup_concurrency require: false type: INT default: 0 min_value: 0 max_value: 100 section: OBSERVER need_restart: false description_en: backup concurrency limit. description_local: observer备份基线的并发度 - name: balancer_log_interval require: false type: TIME default: 1m min_value: 1s max_value: NULL section: LOAD_BALANCE need_restart: false description_en: the time interval between logging the load-balancing tasks statistics. description_local: 负载均衡等后台任务线程打印统计日志的间隔时间 - name: restore_concurrency require: false type: INT default: 0 min_value: 0 max_value: 512 section: OBSERVER need_restart: false description_en: the current work thread num of restore macro block. description_local: 从备份恢复租户数据时最大并发度 - name: micro_block_merge_verify_level require: false type: INT default: 2 min_value: 0 max_value: 3 section: OBSERVER need_restart: false description_en: specify what kind of verification should be done when merging micro block. 0, no verification will be done; 1, verify encoding algorithm, encoded micro block will be read to ensure data is correct; 2, verify encoding and compression algorithm, besides encoding verification, compressed block will be decompressed to ensure data is correct; 3, verify encoding, compression algorithm and lost write protect description_local: 控制合并时宏块的校验级别 - name: bf_cache_miss_count_threshold require: false type: INT default: 100 min_value: 0 max_value: NULL section: CACHE need_restart: false description_en: bf cache miss count threshold, 0 means disable bf cache description_local: 用于控制bloomfilter cache的触发次数，当宏块未命中次数达到这个值时，给创建bloomfilter缓存。0表示关闭。 - name: weak_read_version_refresh_interval require: false type: TIME default: 50ms min_value: 0ms max_value: NULL section: OBSERVER need_restart: false description_en: the time interval to refresh cluster weak read version description_local: 弱一致性读版本号的刷新周期，影响弱一致性读数据的延时；值为0时，表示不再刷新弱一致性读版本号，不提供单调读功能 - name: large_query_worker_percentage require: false type: DOUBLE default: 30 min_value: 0 max_value: 100 section: TENANT need_restart: false description_en: the percentage of the workers reserved to serve large query request. description_local: 预留给大查询处理的工作线程百分比 - name: clog_transport_compress_all require: false type: BOOL default: false min_value: NULL max_value: NULL section: TRANS need_restart: false description_en: If this option is set to true, use compression for clog transport. The default is false(no compression) description_local: 事务日志传输时是否压缩 - name: flush_log_at_trx_commit require: false type: INT default: 1 min_value: 0 max_value: 2 section: TRANS need_restart: false description_en: 0 means commit transactions without waiting clog write to buffer cache, 1 means commit transactions after clog flush to disk, 2 means commit transactions after clog write to buffer cache description_local: 事务提交时写事务日志策略。0表示不等待日志写入缓冲区，1表示等待日志写入磁盘，2表示等待日志写入缓冲区而不等落盘 - name: global_major_freeze_residual_memory require: false type: INT default: 40 min_value: 1 max_value: 99 section: OBSERVER need_restart: false description_en: post global major freeze when observer memsotre free memory(plus memory hold by frozen memstore and blockcache) reach this limit. limit calc by memory_limit * (1 - system_memory_percentage/100) * global_major_freeze_residual_memory/100 description_local: 当剩余内存小于这个百分比时，触发全局冻结 - name: enable_sql_audit require: false type: BOOL default: true min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether SQL audit is turned on. description_local: SQL审计功能开关 - name: merger_switch_leader_duration_time require: false type: TIME default: 3m min_value: 0s max_value: 30m section: ROOT_SERVICE need_restart: false description_en: switch leader duration time for daily merge. description_local: 合并时，批量切主的时间间隔 - name: enable_record_trace_log require: false type: BOOL default: false min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether to always record the trace log. description_local: 是否记录追踪日志 - name: sys_bkgd_migration_change_member_list_timeout require: false type: TIME default: 1h min_value: 0s max_value: 24h section: OBSERVER need_restart: false description_en: the timeout for migration change member list retry. description_local: 副本迁移时变更Paxos成员组操作的超时时间 - name: rootservice_list require: false type: STRING_LIST default: min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: a list of servers which contains rootservice description_local: RootService及其副本所在的机器列表 - name: enable_syslog_wf require: false type: BOOL default: true min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether any log message with a log level higher than WARN would be printed into a separate file with a suffix of wf description_local: 是否把WARN以上级别的系统日志打印到一个单独的日志文件中 - name: global_index_build_single_replica_timeout require: false type: TIME default: 48h min_value: 1h max_value: NULL section: ROOT_SERVICE need_restart: false description_en: build single replica task timeout when rootservice schedule to build global index. description_local: 建全局索引时，每个副本构建的超时时间 - name: memstore_limit_percentage require: false type: INT default: 50 min_value: 1 max_value: 99 section: TENANT need_restart: false description_en: used in calculating the value of MEMSTORE_LIMIT description_local: 租户用于memstore的内存占其总可用内存的百分比 - name: minor_deferred_gc_time require: false type: TIME default: 0s min_value: 0s max_value: 24h section: OBSERVER need_restart: false description_en: sstable deferred gc time after merge description_local: 合并之后SSTable延迟回收间隔 - name: data_disk_usage_limit_percentage require: false type: INT default: 90 min_value: 50 max_value: 100 section: OBSERVER need_restart: false description_en: the safe use percentage of data disk description_local: 数据文件最大可以写入的百分比，超过这个阈值后，禁止数据迁入 - name: enable_perf_event require: false type: BOOL default: true min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether to enable perf event feature. description_local: perf event调试特性开关 - name: obconfig_url require: false type: STRING default: min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: URL for OBConfig service description_local: OBConfig服务的URL地址 - name: system_memory type: CAPACITY default: 30G min_value: 0M max_value: NULL section: OBSERVER need_restart: false description_en: the memory reserved for internal use which cannot be allocated to any outer-tenant, and should be determined to guarantee every server functions normally. description_local: 系统预留内存大小，不能分配给普通租户使用 - name: cpu_quota_concurrency require: false type: DOUBLE default: 4 min_value: 1 max_value: 10 section: TENANT need_restart: false description_en: max allowed concurrency for 1 CPU quota description_local: 租户每个CPU配额允许的最大并发数 - name: zone_merge_order require: false type: STRING default: min_value: NULL max_value: NULL section: DAILY_MERGE need_restart: false description_en: the order of zone start merge in daily merge description_local: 轮转合并的时候，多个Zone的顺序。不指定的时候，由系统自动决定。 - name: backup_recovery_window require: false type: TIME default: 0 min_value: 0 max_value: NULL section: OBSERVER need_restart: false description_en: backup expired day limit, 0 means not expired description_local: 恢复窗口大小 - name: default_row_format require: false type: STRING default: compact min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: default row format in mysql mode description_local: MySQL模式下，建表时使用的默认行格式 - name: token_reserved_percentage require: false type: DOUBLE default: 30 min_value: 0 max_value: 100 section: TENANT need_restart: false description_en: specifies the amount of token increase allocated to a tenant based on his consumption from the last round (without exceeding his upper limit). description_local: 控制租户CPU调度中每次预留多少比例的空闲token数给租户 - name: stack_size require: false type: CAPACITY default: 1M min_value: 512K max_value: 20M section: OBSERVER need_restart: true description_en: the size of routine execution stack description_local: 程序函数调用栈大小 - name: balancer_idle_time require: false type: TIME default: 5m min_value: 10s max_value: NULL section: LOAD_BALANCE need_restart: false description_en: the time interval between the schedules of the partition load-balancing task. description_local: 负载均衡等后台任务线程空闲时的唤醒间隔时间 - name: memory_limit require: false type: CAPACITY default: 0 min_value: NULL max_value: NULL modify_limit: decrease section: OBSERVER need_restart: false description_en: the size of the memory reserved for internal use(for testing purpose) description_local: 可用总内存大小。用于调试，不要设置。 - name: __min_full_resource_pool_memory require: true type: INT default: 2147483648 min_value: 268435456 max_value: NULL need_restart: false description_en: the minimum memory limit of the resource pool description_local: 资源池最小内存限制 - name: virtual_table_location_cache_expire_time require: false type: TIME default: 8s min_value: 1s max_value: NULL section: LOCATION_CACHE need_restart: false description_en: expiration time for virtual table location info in partiton location cache. description_local: 虚拟表的位置信息缓存过期时间 - name: ssl_external_kms_info require: false type: STRING default: min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: when using the external key management center for ssl, this parameter will store some key management information description_local: 配置ssl使用的主密钥管理服务 - name: enable_sql_operator_dump require: false type: BOOL default: true min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether sql operators (sort/hash join/material/window function/interm result/...) allowed to write to disk description_local: 是否允许SQL处理过程的中间结果写入磁盘以释放内存 - name: enable_rich_error_msg require: false type: BOOL default: false min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether add ip:port, time and trace id to user error message. description_local: 是否在客户端消息中添加服务器地址、时间、追踪ID等调试信息 - name: log_archive_concurrency require: false type: INT default: 0 min_value: 0 max_value: NULL section: OBSERVER need_restart: false description_en: concurrency for log_archive_sender and log_archive_spiter description_local: 日志归档并发度 - name: server_balance_disk_tolerance_percent require: false type: INT default: 1 min_value: 1 max_value: 100 section: LOAD_BALANCE need_restart: false description_en: specifies the tolerance (in percentage) of the unbalance of the disk space utilization among all servers. The average disk space utilization is calculated by dividing the total space by the number of servers. server balancer will start a rebalancing task when the deviation between the average usage and some server load is greater than this tolerance description_local: 节点负载均衡策略中，磁盘资源不均衡的容忍度 - name: user_tab_col_stat_cache_priority require: false type: INT default: 1 min_value: 1 max_value: NULL section: CACHE need_restart: false description_en: user tab col stat cache priority description_local: 统计数据缓存在缓存系统中的优先级 - name: recyclebin_object_expire_time require: false type: TIME default: 0s min_value: 0s max_value: NULL section: ROOT_SERVICE need_restart: false description_en: recyclebin object expire time, default 0 that means auto purge recyclebin off. description_local: 回收站对象的有效期，超过有效的对象将被回收；0表示关闭回收功能； - name: minor_warm_up_duration_time require: false type: TIME default: 30s min_value: 0s max_value: 60m section: OBSERVER need_restart: false description_en: warm up duration time for minor freeze. description_local: 小合并产生新转储文件的预热时间 - name: migrate_concurrency require: false type: INT default: 10 min_value: 0 max_value: 64 section: OBSERVER need_restart: false description_en: set concurrency of migration, set upper limit to migrate_concurrency and set lower limit to migrate_concurrency/2 description_local: 控制内部数据迁移的并发度 - name: redundancy_level require: false type: STRING default: NORMAL min_value: NULL max_value: NULL section: SSTABLE need_restart: false description_en: EXTERNAL, use extrernal redundancy; NORMAL, tolerate one disk failure, HIGH tolerate two disk failure if disk count is enough description_local: OB内置本地磁盘RAID特性。暂勿使用 - name: rootservice_memory_limit require: false type: CAPACITY default: 2G min_value: 2G max_value: NULL section: OBSERVER need_restart: false description_en: max memory size which can be used by rs tenant description_local: RootService最大内存限制 - name: plan_cache_low_watermark require: false type: CAPACITY default: 1500M min_value: NULL max_value: NULL section: TENANT need_restart: false description_en: memory usage at which plan cache eviction will be stopped. description_local: 执行计划缓存占用内存低于该阈值时将停止淘汰 - name: trx_2pc_retry_interval require: false type: TIME default: 100ms min_value: 1ms max_value: 5000ms section: TRANS need_restart: false description_en: the time interval between the retries in case of failure during a transactions two-phase commit phase description_local: 两阶段提交失败时候自动重试的间隔 - name: global_write_halt_residual_memory require: false type: INT default: 30 min_value: 1 max_value: 99 section: OBSERVER need_restart: false description_en: disable write to memstore when observer memstore free memory(plus memory hold by blockcache) lower than this limit, description_local: 当全局剩余内存小于这个百分比时，暂停普通租户写入（sys租户不受影响） - name: cpu_count require: false type: INT default: 0 min_value: 0 max_value: NULL section: OBSERVER need_restart: true description_en: the number of CPUs in the system. If this parameter is set to zero, the number will be set according to sysconf; otherwise, this parameter is used. description_local: 系统CPU总数，如果设置为0，将自动检测 - name: max_syslog_file_count require: false type: INT default: 0 min_value: 0 max_value: NULL section: OBSERVER need_restart: false description_en: specifies the maximum number of the log files that can co-exist before the log file recycling kicks in. Each log file can occupy at most 256MB disk space. When this value is set to 0, no log file will be removed. description_local: 系统日志自动回收复用时，最多保留多少个。值0表示不自动清理。 - name: appname require: false type: STRING default: obcluster min_value: NULL max_value: NULL section: OBSERVER need_redeploy: true description_en: Name of the cluster description_local: 本OceanBase集群名 - name: use_large_pages require: false type: STRING default: false min_value: NULL max_value: NULL section: OBSERVER need_restart: true description_en: used to manage the databases use of large pages, values are false, true, only description_local: 控制内存大页的行为,"true"表示在操作系统开启内存大页并且有空闲大页时，数据库总是申请内存大页，否则申请普通内存页, "false"表示数据库不使用大页, "only"表示数据库总是分配大页 - name: dtl_buffer_size require: false type: CAPACITY default: 64K min_value: 4K max_value: 2M section: OBSERVER need_restart: false description_en: buffer size for DTL description_local: SQL数据传输模块使用的缓存大小 - name: server_balance_critical_disk_waterlevel require: false type: INT default: 80 min_value: 0 max_value: 100 section: LOAD_BALANCE need_restart: false description_en: disk water level to determine server balance strategy description_local: 磁盘水位线超过该阈值时，负载均衡策略将倾向于优先考虑磁盘均衡 - name: location_fetch_concurrency require: false type: INT default: 20 min_value: 1 max_value: 1000 section: LOCATION_CACHE need_restart: false description_en: the maximum number of the tasks which fetch the partition location information concurrently. description_local: 位置缓存信息刷新的最大并发度 - name: enable_async_syslog require: false type: BOOL default: true min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: specifies whether use async syslog description_local: 是否启用系统日志异步写 - name: clog_sync_time_warn_threshold require: false type: TIME default: 1s min_value: 1ms max_value: 10000ms section: TRANS need_restart: false description_en: the time given to the commit log synchronization between a leader and its followers before a warning message is printed in the log file. description_local: 事务日志同步耗时告警阈值，同步耗时超过该值产生WARN日志 - name: location_cache_cpu_quota require: false type: DOUBLE default: 5 min_value: 0 max_value: 10 section: TENANT need_restart: false description_en: the number of vCPUs allocated for the requests regarding location info of the core tables. description_local: 位置缓存模块使用的CPU配额 - name: bf_cache_priority require: false type: INT default: 1 min_value: 1 max_value: NULL section: CACHE need_restart: false description_en: bloomfilter cache priority description_local: 布隆过滤器占用缓存的优先级 - name: merger_check_interval require: false type: TIME default: 10m min_value: 10s max_value: 60m section: DAILY_MERGE need_restart: false description_en: the time interval between the schedules of the task that checks on the progress of MERGE for each zone. description_local: 合并状态检查线程的调度间隔 - name: enable_rootservice_standalone require: false type: BOOL default: false min_value: NULL max_value: NULL section: ROOT_SERVICE need_restart: false description_en: specifies whether the SYS tenant is allowed to occupy an observer exclusively, thus running in the standalone mode. description_local: 是否让系统租户和RootService独占observer节点 - name: px_workers_per_cpu_quota require: false type: INT default: 10 min_value: 0 max_value: 20 section: TENANT need_restart: false description_en: the ratio between the number of system allocated px workers vs the maximum number of threads that can be scheduled concurrently. description_local: 并行执行工作线程数的比例 - name: large_query_threshold require: false type: TIME default: 100ms min_value: 1ms max_value: NULL section: TENANT need_restart: false description_en: threshold for execution time beyond which a request may be paused and rescheduled as large request description_local: 一个查询执行时间超过该阈值会被判断为大查询，执行大查询调度策略 - name: sys_bkgd_net_percentage require: false type: INT default: 60 min_value: 0 max_value: 100 section: OBSERVER need_restart: false description_en: the net percentage of sys background net. description_local: 后台系统任务可占用网络带宽百分比 - name: fuse_row_cache_priority require: false type: INT default: 1 min_value: 1 max_value: NULL section: CACHE need_restart: false description_en: fuse row cache priority description_local: 融合行缓存在缓存系统中的优先级 - name: rpc_timeout require: false type: TIME default: 2s min_value: NULL max_value: NULL section: RPC need_restart: false description_en: the time during which a RPC request is permitted to execute before it is terminated description_local: 集群内部请求的超时时间 - name: tenant_task_queue_size require: false type: INT default: 65536 min_value: 1024 max_value: NULL section: OBSERVER need_restart: false description_en: the size of the task queue for each tenant. description_local: 每个租户的请求队列大小 - name: resource_soft_limit require: false type: INT default: 50 min_value: 1 max_value: 10000 section: LOAD_BALANCE need_restart: false description_en: Used along with resource_hard_limit in unit allocation. If server utilization is less than resource_soft_limit, a policy of best fit will be used for unit allocation; otherwise, a least loadpolicy will be employed. Ultimately,system utilization should not be large than resource_hard_limit. description_local: 当所有节点的资源水位低于该阈值时，不执行负载均衡 - name: plan_cache_evict_interval require: false type: TIME default: 1s min_value: 0s max_value: NULL section: TENANT need_restart: false description_en: time interval for periodic plan cache eviction. description_local: 执行计划缓存的淘汰间隔 - name: server_balance_cpu_mem_tolerance_percent require: false type: INT default: 5 min_value: 1 max_value: 100 section: LOAD_BALANCE need_restart: false description_en: specifies the tolerance (in percentage) of the unbalance of the cpu/memory utilization among all servers. The average cpu/memory utilization is calculated by dividing the total cpu/memory by the number of servers. server balancer will start a rebalancing task when the deviation between the average usage and some server load is greater than this tolerance description_local: 节点负载均衡策略中，CPU和内存资源不均衡的容忍度 - name: autoinc_cache_refresh_interval require: false type: TIME default: 3600s min_value: 100ms max_value: NULL section: OBSERVER need_restart: false description_en: auto-increment service cache refresh sync_value in this interval description_local: 自动刷新自增列值的时间间隔 - name: all_server_list require: false type: STRING default: min_value: NULL max_value: NULL section: LOCATION_CACHE need_restart: false description_en: all server addr in cluster description_local: 集群中所有机器的列表，不建议人工修改 - name: enable_rebalance require: false type: BOOL default: true min_value: NULL max_value: NULL section: LOAD_BALANCE need_restart: false description_en: specifies whether the partition load-balancing is turned on. description_local: 自动负载均衡开关 - name: internal_sql_execute_timeout require: false type: TIME default: 30s min_value: 1000us max_value: 10m section: OBSERVER need_restart: false description_en: the number of microseconds an internal DML request is permitted to execute before it is terminated. description_local: 系统内部SQL请求的超时时间 - name: user_row_cache_priority require: false type: INT default: 1 min_value: 1 max_value: NULL section: CACHE need_restart: false description_en: user row cache priority description_local: 基线数据行缓存在缓存系统中的优先级 - name: server_permanent_offline_time require: false type: TIME default: 3600s min_value: 20s max_value: NULL section: ROOT_SERVICE need_restart: false description_en: the time interval between any two heartbeats beyond which a server is considered to be permanently offline. description_local: 节点心跳中断多久后认为其被“永久下线”，“永久下线”的节点上的数据副本需要被自动补足 - name: schema_history_expire_time require: false type: TIME default: 7d min_value: 1m max_value: 30d section: OBSERVER need_restart: false description_en: the hour of expire time for schema history description_local: 元数据历史数据过期时间 - name: datafile_disk_percentage require: false type: INT min_value: 0 max_value: 99 modify_limit: decrease section: SSTABLE need_restart: false description_en: the percentage of disk space used by the data files. description_local: data_dir所在磁盘将被OceanBase系统初始化用于存储数据，本配置项表示占用该磁盘总空间的百分比 - name: default_compress_func require: false type: STRING default: zstd_1.3.8 min_value: NULL max_value: NULL section: OBSERVER need_restart: false description_en: default compress function name for create new table description_local: MySQL模式下，建表时使用的默认压缩算法 - name: memory_chunk_cache_size require: false type: CAPACITY default: 0M min_value: 0M max_value: NULL section: OBSERVER need_restart: false description_en: the maximum size of memory cached by memory chunk cache. description_local: 内存分配器缓存的内存块容量。值为0的时候表示系统自适应。 - name: ob_event_history_recycle_interval require: false type: TIME default: 7d min_value: 1d max_value: 180d section: ROOT_SERVICE need_restart: false description_en: the time to recycle event history. description_local: OB事件表中事件条目的保存期限 - name: enable_ddl require: false type: BOOL default: true min_value: NULL max_value: NULL section: ROOT_SERVICE need_restart: false description_en: specifies whether DDL operation is turned on. description_local: 是否允许执行DDL - name: balance_blacklist_failure_threshold require: false type: INT default: 5 min_value: 0 max_value: 1000 section: LOAD_BALANCE need_restart: false description_en: a balance task failed count to be putted into blacklist description_local: 副本迁移等后台任务连续失败超过该阈值后，将被放入黑名单 - name: system_trace_level require: false type: INT default: 1 min_value: 0 max_value: 2 section: OBSERVER need_restart: false description_en: system trace log level, 0:none, 1:standard, 2:debug. description_local: 系统追踪日志的日志打印级别 - name: wait_leader_batch_count require: false type: INT default: 1024 min_value: 128 max_value: 5000 section: ROOT_SERVICE need_restart: false description_en: leader batch count everytime leader coordinator wait. description_local: RootService发送切主命令的批次大小 - name: trace_log_sampling_interval require: false type: TIME default: 10ms min_value: 0ms max_value: NULL section: OBSERVER need_restart: false description_en: the time interval for periodically printing log info in trace log. When force_trace_log is set to FALSE, for each time interval specifies by sampling_trace_log_interval, logging info regarding ‘slow query’ and ‘white list’ will be printed out. description_local: 追踪日志的采样间隔，当force_trace_log关闭的时候生效 - name: proxyro_password require: false type: STRING default: '' min_value: NULL max_value: NULL need_restart: false description_en: password of observer proxyro user description_local: proxyro用户的密码 - name: root_password require: false type: STRING default: '' min_value: NULL max_value: NULL need_restart: false description_en: password of observer root user description_local: sys租户root用户的密码 - name: log_disk_percentage require: false type: INT default: 0 min_value: 0 max_value: 99 description_en: the percentage of disk space used by the clog files. description_local: Redo 日志占用其所在磁盘总空间的百分比。 - name: log_disk_size require: false type: CAPACITY default: 0 min_value: 0M max_value: NULL description_en: the size of disk space used by the clog files. description_local: Redo 日志磁盘的大小。 # todo： 等文档更新 - name: sql_login_thread_count require: false type: INT default: 0 min_value: 0 max_value: 32 section: OBSERVER need_restart: false need_redeploy: false description_en: 'the number of threads for sql login request. Range: [0, 32] in integer, 0 stands for use default thread count defined in TG.the default thread count for login request in TG is normal:6 mini-mode:2' description_local: '' - name: tcp_keepcnt require: false type: INT default: 10 min_value: 1 max_value: NULL section: OBSERVER need_restart: false need_redeploy: false description_en: 'The maximum number of keepalive probes TCP should send before dropping the connection. Take effect for new established connections. Range: [1,+∞)' description_local: 关闭一个非活跃连接之前的最大重试次数。 - name: tcp_keepintvl require: false type: TIME default: 6s min_value: 1s max_value: NULL section: OBSERVER need_restart: false need_redeploy: false description_en: 'The time (in seconds) between individual keepalive probes. Take effect for new established connections. Range: [1s, +∞]' description_local: 开启客户端连接的探活机制后，前后两次探测之间的时间间隔，单位为秒。 - name: tcp_keepidle require: false type: TIME default: 7200s min_value: 1s max_value: NULL section: OBSERVER need_restart: false need_redeploy: false description_en: 'The time (in seconds) the connection needs to remain idle before TCP starts sending keepalive probe. Take effect for new established connections. Range: [1s, +∞]' description_local: 客户端连接上服务器后，如果没有数据发送，多久后会发送 Keepalive 探测分组，单位为秒。 - name: enable_tcp_keepalive require: false type: BOOL default: true min_value: NULL max_value: NULL section: OBSERVER need_restart: false need_redeploy: false description_en: enable TCP keepalive for the TCP connection of sql protocol. Take effect for new established connections. description_local: 开启或关闭客户端连接的探活机制。 - name: ob_ratelimit_stat_period require: false type: TIME default: 1s min_value: 100ms max_value: NULL section: OBSERVER need_restart: false need_redeploy: false description_en: "the time interval to update observer's maximum bandwidth to a certain region. " description_local: OBServer 计算和更新最大带宽的时间间隔。 - name: enable_ob_ratelimit require: false type: BOOL default: false min_value: NULL max_value: NULL section: OBSERVER need_restart: false need_redeploy: false description_en: enable ratelimit between regions for RPC connection. description_local: 开启或关闭客户端连接的探活机制。 - name: schema_history_recycle_interval require: false type: TIME default: 10m min_value: 0s max_value: NULL section: LOAD_BALANCE need_restart: false need_redeploy: false description_en: 'the time interval between the schedules of schema history recyle task. Range: [0s, +∞)' description_local: 系统内部执行 schema 多版本记录回收任务的时间间隔。 - name: backup_data_file_size require: false type: CAPACITY default: 4G min_value: 512M max_value: 4G section: OBSERVER need_restart: false need_redeploy: false description_en: 'backup data file size. Range: [512M, 4G] in integer' description_local: 备份数据文件的容量。 - name: data_storage_error_tolerance_time require: false type: TIME default: 300s min_value: 10s max_value: 7200s section: OBSERVER need_restart: false need_redeploy: false description_en: time to tolerate disk read failure, after that, the disk status will be set error. Range [10s,7200s]. The default value is 300s description_local: 数据盘状态设为 ERROR 状态的容忍时间。 - name: data_storage_warning_tolerance_time require: false type: TIME default: 30s min_value: 10s max_value: 300s section: OBSERVER need_restart: false need_redeploy: false description_en: time to tolerate disk read failure, after that, the disk status will be set warning. Range [10s,300s]. The default value is 30s description_local: 数据盘状态设为 WARNING 状态的容忍时间。 - name: index_block_cache_priority require: false type: INT default: 10 min_value: 1 max_value: NULL section: CACHE need_restart: false need_redeploy: false description_en: index cache priority. Range:[1, ) description_local: Tablet 映射缓存优先级。 - name: opt_tab_stat_cache_priority require: false type: INT default: 1 min_value: 1 max_value: NULL section: CACHE need_restart: false need_redeploy: false description_en: tab stat cache priority. Range:[1, ) description_local: 统计信息缓存优先级。 - name: tablet_ls_cache_priority require: false type: INT default: 1000 min_value: 1 max_value: NULL section: CACHE need_restart: false need_redeploy: false description_en: tablet ls cache priority. Range:[1, ) description_local: 元数据索引微块缓存优先级。 - name: location_cache_refresh_sql_timeout require: false type: TIME default: 1s min_value: 1ms max_value: NULL section: LOCATION_CACHE need_restart: false need_redeploy: false description_en: 'The timeout used for refreshing location cache by SQL. Range: [1ms, +∞)' description_local: 通过 SQL 刷新 location_cache 的超时时间。 - name: location_cache_refresh_rpc_timeout require: false type: TIME default: 500ms min_value: 1ms max_value: NULL section: LOCATION_CACHE need_restart: false need_redeploy: false description_en: 'The timeout used for refreshing location cache by RPC. Range: [1ms, +∞)' description_local: 通过 RPC 刷新 location_cache 的超时时间。 - name: tablet_meta_table_check_interval require: false type: TIME default: 30m min_value: 1m max_value: NULL section: ROOT_SERVICE need_restart: false need_redeploy: false description_en: 'the time interval that observer compares tablet meta table with local ls replica info and make adjustments to ensure the correctness of tablet meta table. Range: [1m,+∞)' description_local: DBA_OB_TABLET_REPLICAS/CDB_OB_TABLET_REPLICAS 视图的后台巡检线程的检查间隔。 - name: ls_meta_table_check_interval require: false type: TIME default: 1s min_value: 1ms max_value: NULL section: ROOT_SERVICE need_restart: false need_redeploy: false description_en: 'the time interval that observer compares ls meta table with local ls replica info and make adjustments to ensure the correctness of ls meta table. Range: [1ms,+∞)' description_local: DBA_OB_LS_LOCATIONS/CDB_OB_LS_LOCATIONS 视图的后台巡检线程的检查间隔。 - name: tablet_meta_table_scan_batch_count require: false type: INT default: 999 min_value: 1 max_value: 65536 section: ROOT_SERVICE need_restart: false need_redeploy: false description_en: the number of tablet replica info that will be read by each request on the tablet-related system tables during procedures such as load-balancing, daily merge, election and etc. Range:(0,65536] description_local: Tablet meta table 迭代器使用过程中在内存里缓存的 Tablet 数量。 - name: rdma_io_thread_count require: false type: INT default: 0 min_value: 0 max_value: 8 section: OBSERVER need_restart: true need_redeploy: false description_en: 'the number of RDMA I/O threads for Libreasy. Range: [0, 8] in integer, 0 stands for RDMA being disabled.' description_local: Libreasy 的 RDMA I/O 线程数。 - name: production_mode require: true type: BOOL default: true min_value: NULL max_value: NULL section: OBSERVER need_restart: false need_redeploy: false description_en: Production mode switch, default True. Adjust the memory_limit and __min_full_resource_pool_memory The lower bound of memory is adjusted to 16G and 2147483648 description_local: 生产模式开关, 默认开启。开启后调整memory limit 和 __min_full_resource_pool_memory 下界调整为 16G 和 2147483648