2. ClustrixDB 文件/参数说明

一、日志/data/clustrix/log/query.log

记录节点慢SQL/错误SQL/DDL 等信息，节点分开记录

Each entry in the query.log is categorized as one of these types. Specific logging for each query type is controlled by the global or session variable indicated.

Query Type	Description
ALTER CLUSTER	Changes made to your cluster via the ALTER CLUSTER command are always logged to the query.log automatically. This logging is not controlled by a global variable.
BAD	The query reads more rows than necessary to return the expected results. This may indicate a bad plan or missing index. Logging of BAD queries is not enabled by default (session_log_bad_queries).
DDL	The query is DDL (i.e. schema change such as CREATE, DROP, ALTER), or a SET GLOBAL or SESSION command. All DDL queries are initially logged by default (session_log_ddl).
SLOW	Query execution time exceeded the threshold specified by the variable session_log_slow_threshold_ms.
SQLERR	These database errors are things such as syntax errors, timeout notifications, and permission issues. All SQLERR queries will be logged by default (session_log_error_queries).

These are the variables that control query and user logging. The defaults shown are generally acceptable for most installations.

Name	Description	Default Value	Session Variable
session_log_bad_queries	Log BAD queries to the query.log	false	‌
session_log_ddl	Log DDL statements to query.log	true
session_log_error_queries	Log ERROR statements to query.log	true
session_log_slow_queries	Log SLOW statements to query.log	true
session_log_slow_threshold_ms	Query duration threshold in milliseconds before logging this query	10000	‌
session_log_users	Log LOGIN/LOGOUT to user.log	false

MySQL [(none)]> show variables like '%session_log%';

+--------------------------------+--------+

| Variable_name                  | Value  |

+--------------------------------+--------+

| session_log_bad_queries        | false  |

| session_log_bad_read_ratio     | 100    |

| session_log_bad_read_threshold | 4000   |

| session_log_ddl                | true   |

| session_log_error_queries      | true   |

| session_log_slow_queries       | true   |

| session_log_slow_threshold_ms  | 100000 |

| session_log_users              | false  |

+--------------------------------+--------+

8 rows in set (0.01 sec)

修改慢SQL时间，超过1s的SQL都被记录

MySQL [(none)]> set global session_log_slow_threshold_ms=1000;
Query OK, 0 rows affected (0.01 sec)

MySQL [(none)]> show global variables like 'session_log_slow_threshold_ms';
+-------------------------------+-------+
| Variable_name | Value |
+-------------------------------+-------+
| session_log_slow_threshold_ms | 1000 |
+-------------------------------+-------+

查看慢SQL日志（集群所有的操作）：

MySQL [(none)]>  select * from tail_query_log order by timestamp desc limit 10;

二、 日志 /data/clustrix/log/user.log

监控用户登录登出状态，单个节点文件记录单个节点信息，查看集群，system.tail_user_log表中查询。

默认关闭监控用户状态

MySQL [system]> show variables like 'session_log_users';
+-------------------+-------+
| Variable_name | Value |
+-------------------+-------+
| session_log_users | false |
+-------------------+-------+
1 row in set (0.01 sec)

MySQL [system]> set global session_log_users=1;
Query OK, 0 rows affected (0.51 sec)

MySQL [system]> show variables like 'session_log_users';
+-------------------+-------+
| Variable_name | Value |
+-------------------+-------+
| session_log_users | true |
+-------------------+-------+

查看集群用户登录登出状态

MySQL [system]> select * from system.tail_user_log;

+--------+----------------------------+------+----------+---------+---------------------------------------------------------------------------------------------------------------+----------+
| nodeid | timestamp | num | hostname | command | message | repeated |
+--------+----------------------------+------+----------+---------+---------------------------------------------------------------------------------------------------------------+----------+
| 1 | 2019-11-26 07:22:23.396171 | 0 | nid | NULL | 1 ip-10-1-3-88.cn-northwest-1.compute.internal clxnode: USER SID:16385 db="system" user=root@localhost LOGOUT | 0 |
| 1 | 2019-11-26 07:22:24.371812 | 1 | nid | NULL | 1 ip-10-1-3-88.cn-northwest-1.compute.internal clxnode: USER SID:17409 db=#undef user=root@localhost LOGIN | 0 |
| 1 | 2019-11-26 07:22:54.216004 | 2 | nid | NULL | 1 ip-10-1-3-88.cn-northwest-1.compute.internal clxnode: USER SID:18433 db=#undef user=root@localhost LOGIN | 0 |
| 1 | 2019-11-26 07:23:05.066316 | 3 | nid | NULL | 1 ip-10-1-3-88.cn-northwest-1.compute.internal clxnode: USER SID:18433 db=#undef user=root@localhost LOGOUT | 0 |
| 2 | 2019-11-26 07:23:07.523621 | 0 | nid | NULL | 2 ip-10-1-3-242.cn-northwest-1.compute.internal clxnode: USER SID:6146 db="system" user=root@localhost LOGOUT | 0 |
| 2 | 2019-11-26 07:23:08.049621 | 1 | nid | NULL | 2 ip-10-1-3-242.cn-northwest-1.compute.internal clxnode: USER SID:7170 db=#undef user=root@localhost LOGIN | 0 |
+--------+----------------------------+------+----------+---------+---------------------------------------------------------------------------------------------------------------+----------+
6 rows in set (0.00 sec)

三、Rebalance 配置

查看rebalance状态

sql> select * from system.rebalancer_activity_log order by started desc limit 10; 

提高rebalance性能

sql> set global rebalancer_rebalance_task_limit = 8;
sql> set global rebalancer_vdev_task_limit = 4;
sql> set global task_rebalancer_rebalance_distribution_interval_ms = 5000;
sql> set global task_rebalancer_rebalance_interval_ms = 5000;

如果出现负载过高

sql> set global rebalancer_rebalance_task_limit = default;

sql> set global rebalancer_vdev_task_limit = default;
sql> set global task_rebalancer_rebalance_distribution_interval_ms = default;
sql> set global task_rebalancer_rebalance_interval_ms = default;

四、查看那些配置不是默认值

sql> select * from system.global_variable_definitions where current_value != default_value;

五、监控SQL

列出消耗CPU最大的3个历史SQL

SELECT nodeid, exec_count, exec_ms, exec_ms/exec_count as avg_ms, left(statement,100)
     FROM system.qpc_queries
     ORDER BY exec_ms desc
     LIMIT 3;

列出过去24小时内最频繁运行的100个SQL

sql> SELECT query_key,

             min(rank),
             max(rank),
             database,
             left(statement,100),
             sum(exec_count) as calc_exec_count,
             round(avg(avg_rows_read)) as calc_avg_rows_read,
             round(avg(avg_exec_ms)) as calc_avg_exec_ms
     FROM    clustrix_statd.qpc_history
     WHERE   timestamp BETWEEN (now() - interval 24 hour) AND now()
     AND     database !='clustrix_statd'
     GROUP BY query_key
     ORDER BY calc_exec_count     DESC,
              calc_avg_rows_read  DESC
     LIMIT 100;

列出过去24小时内返回行数最多的100个SQL

sql> SELECT query_key,
            min(rank),
            max(rank),
            database,
            left(statement,100),
            sum(exec_count) as calc_exec_count,
            round(avg(avg_rows_read)) as calc_avg_rows_read,
            round(avg(avg_exec_ms)) as calc_avg_exec_ms
     FROM   clustrix_statd.qpc_history
     WHERE  timestamp BETWEEN (now() - interval 24 hour) AND now()
     AND    database !='clustrix_statd'
     GROUP BY query_key
     ORDER BY calc_avg_rows_read DESC,
              calc_exec_count DESC
     LIMIT 100;

列出过去24小时内返回运行时间最长的100个SQL

sql> SELECT query_key,
            min(rank),
            max(rank),
            database,
            left(statement,100),
            sum(exec_count) as calc_exec_count,
            round(avg(avg_rows_read)) as calc_avg_rows_read,
            round(avg(avg_exec_ms)) as calc_avg_exec_ms
     FROM   clustrix_statd.qpc_history
     WHERE  timestamp BETWEEN (now() - interval 24 hour) AND now()
     AND    database !='clustrix_statd'
     GROUP BY query_key
     ORDER BY calc_avg_exec_ms  DESC,
              calc_exec_count   DESC
     LIMIT 100;

六、system's tables

MySQL [(none)]> use system;
Database changed
MySQL [system]> show tables;
+-----------------------------------+
| Tables_in_system |
+-----------------------------------+
| activity |
| alerts_intervals |
| alerts_messages |
| alerts_parameters |
| alerts_subscriptions |
| alter_progress |
| autoinc_sequences |
| backups |
| backup_masters |
| backup_status |
| backup_tables |
| barriers |
| base_allocators |
| bigc_state |
| binlogs |
| binlog_commits |
| binlog_commits_segments |
| binlog_ignore_databases |
| binlog_ignore_tables |
| binlog_log_databases |
| binlog_log_tables |
| binlog_segments |
| bm_latch_waits |
| bm_stats |
| broadcast_nodes |
| check_constraints |
| cluster_session_stats |
| cluster_session_variables |
| columns |
| constraints |
| containers |
| container_stats |
| container_truncates |
| container_type_codes |
| cpm_history |
| cpm_info |
| cpu_activity |
| cpu_allocations |
| cpu_load |
| databases |
| debugpoints |
| deferred_foreign_keys |
| deferred_foreign_key_columns |
| definers |
| device_containers |
| device_containers2 |
| device_space_stats |
| disks |
| disk_activity |
| disk_paths |
| dlog_stats |
| dropped_binlogs |
| engines |
| error_codes |
| event_map |
| failpoints |
| fibers |
| flow_control_channels |
| flow_control_peers |
| foreign_keys |
| foreign_key_columns |
| global_stats |
| global_variables |
| global_variables_ignored |
| global_variable_definitions |
| groups |
| gtm_accepter |
| gtm_coord |
| gtm_coord_invocations |
| gtm_ddl |
| gtm_ddl_server |
| gtm_poisoned_invoc |
| gtm_poisoned_trx |
| gtm_repick_accepters |
| gtm_resolver |
| gtm_send_queues |
| hash_distribution_map |
| heaps |
| imported_index_pds |
| index_sizes |
| index_stats |
| init_graph_stats |
| internal_routines |
| internode_latency |
| invocations |
| irp_queues |
| key_caches |
| layercons |
| layers |
| layer_merges |
| license |
| load |
| lockman |
| lockman_holders |
| lockman_victims |
| lpdcache |
| lpds |
| ltm_transactions |
| mdstat |
| media_scanners |
| membership |
| memory_table_replicas |
| missing_pds |
| missing_pd_columns |
| missing_pd_details |
| mounts |
| mvcc_waiters |
| mysql_binlogs |
| mysql_binlog_index |
| mysql_binlog_segments |
| mysql_binlog_stats |
| mysql_binlog_trims |
| mysql_character_sets |
| mysql_collations |
| mysql_db_replication_policy |
| mysql_error_codes |
| mysql_indexed_binlogs |
| mysql_master_status |
| mysql_registered_slaves |
| mysql_repconfig |
| mysql_repslave_svars |
| mysql_repstate |
| mysql_repstate_until |
| mysql_sessions |
| mysql_slave_connection_status |
| mysql_slave_db_replication_policy |
| mysql_slave_driver_status |
| mysql_slave_log_updates |
| mysql_slave_rewrite_db |
| mysql_slave_skip_errors |
| mysql_slave_stats |
| mysql_slave_status |
| mysql_slave_variables |
| mysql_table_replication_policy |
| named_locks |
| networking |
| network_activity |
| nodeinfo |
| nodes |
| objects |
| partitioned_hash_distributions |
| partitions |
| partition_endpoints |
| partition_functions |
| pdcache_raw |
| pdms |
| pds |
| pd_groups |
| pd_group_columns |
| pending_invites |
| periodic_tasks |
| poisoned_barriers |
| privileges |
| problem_nodes |
| processlist |
| processlistfull |
| proc_cpu |
| proc_cpu_rates |
| proc_diskstats |
| proc_diskstat_rates |
| proc_interrupts |
| proc_interrupt_rates |
| proc_meminfo |
| proc_net_dev |
| proc_net_dev_rates |
| proc_softirqs |
| proc_softirq_rates |
| profiled_invocations |
| profiled_plans |
| profiled_statements |
| profiled_til |
| profiled_transactions |
| profiling |
| program_cache |
| protection_log |
| ps |
| public_keys |
| qpc_lru |
| qpc_plans |
| qpc_queries |
| queues |
| queue_readers |
| queue_replays |
| queue_replay_streams |
| queue_replay_waiters |
| queue_status |
| range_hash_distributions |
| rebalancer_activity_log |
| rebalancer_activity_targets |
| rebalancer_copies |
| rebalancer_copy_activity |
| rebalancer_copy_work_queue |
| rebalancer_hash_distributions |
| rebalancer_queued_activity |
| rebalancer_redistributes |
| rebalancer_replicas |
| rebalancer_representations |
| rebalancer_scheduled_replicas |
| rebalancer_slices |
| rebalancer_splits |
| rebalancer_started_activity |
| rebalancer_summary |
| rebalancer_vdevs |
| redistributes |
| relations |
| relation_alters |
| relation_alter_columns |
| relation_build_status |
| relation_sizes |
| replicas |
| replicated_containers |
| replication_checkpoint |
| replication_master_status |
| replica_copies |
| replica_sizes |
| replica_status_codes |
| representations |
| representation_builds |
| representation_columns |
| representation_sizes |
| representation_stats |
| restore_database_builds |
| restore_object_builds |
| ril_cache_p |
| ril_stats |
| routines |
| routine_parameters |
| sequences |
| sequence_state |
| sessions |
| session_call_stacks |
| session_containers |
| session_local_variables |
| session_pdcache |
| session_row_stats |
| sighandlers |
| sighandlers_json |
| signals |
| skiplists |
| slave_row_stats |
| slave_slices |
| slave_slice_readers |
| slave_slice_relations |
| slave_slice_writers |
| slices |
| slice_splits |
| softfailed_devices |
| softfailed_nodes |
| softfailing_containers |
| sqlstates |
| stats |
| strmaps |
| suspended_closures |
| system_users |
| sys_block_devs |
| tables_and_views |
| table_pdms |
| table_replicas |
| table_sizes |
| table_slices |
| tail_clustrix_log |
| tail_nanny_log |
| tail_query_log |
| tail_user_log |
| tail_webui_log |
| tasks |
| task_placement |
| tcp |
| temporary_tables |
| time_zones |
| tracepoints |
| transactions |
| triggers |
| trigger_event |
| trigger_orientation |
| trigger_timing |
| trxstate |
| trxstate_stats |
| underprotected_slices |
| users |
| user_accessible_databases |
| user_accessible_tables |
| user_accessible_triggers |
| user_accessible_users |
| user_acl |
| user_routine_acl |
| vdev_io |
| vdev_stat |
| version_history |
| views |
| virtual_relations |
| virtual_views |
| vmstats |
| wals |
| wal_windows |
+-----------------------------------+
295 rows in set (0.14 sec)