Inspection metrics|V3.3.0| docs|Distributed Database

This topic describes the inspection metrics of OceanBase Diagnostic Tool (obdiag).

OceanBase Database

Note

The issue numbers issue #xx mentioned in this topic are those in obdiag.

Type	Task	Inspection purpose
disk	xfs_repair	Checks whether the output of `dmesg` contains xfs_repair logs. For more information, see issue #451 in GitHub.
disk	disk_full	Checks whether the disk usage reaches the threshold.
disk	clog_abnormal_file	Checks whether the `clog` folder contains files that do not belong to the OceanBase cluster.
disk	sstable_abnormal_file	Checks whether the `sstable` folder contains files that do not belong to the OceanBase cluster.
disk	disk_iops	Checks the disk IOPS.
disk	disk_hole	Checks for disk holes.
system	instruction_set_avx	Checks for CPU flags.
system	instruction_set_avx2	Checks for CPU flags.
system	core_pattern_ctl	Checks kernel.core_pattern.
system	core_pattern	Checks kernel.core_pattern.
system	dependent_software	Checks dependent software. For more information, see Host standardization check items.
system	dependent_software_swapon	Checks dependent software. For more information, see Host standardization check items.
system	clock_source	Checks whether the type of the clock source is tsc.
system	ulimit_parameter	Checks ulimit-related parameters. For more information, see Host standardization check items.
system	parameter	Checks kernel parameters. For more information, see Host standardization check items.
system	parameter_ip_local_port_range	Checks kernel parameters. For more information, see Host standardization check items.
system	parameter_tcp_wmem	Checks kernel parameters. For more information, see Host standardization check items.
system	getenforce	Checks SELinux by using the `getenforce` command. For more information, see Host standardization check items.
system	parameter_tcp_rmem	Checks kernel parameters. For more information, see Host standardization check items.
system	aio	Checks the asynchronous I/O (AIO). For more information, see Host standardization check items.
err_code	find_err_4000	Checks whether the error code `4000` is returned when `enable_sql_audit` is set to `True`.
err_code	find_err_4016	Checks whether the error code `4016` is returned when `enable_sql_audit` is set to `True`.
err_code	find_err_4377	Checks whether the error code `4377` is returned when `enable_sql_audit` is set to `True`.
err_code	find_err_4108	Checks whether the error code `4108` is returned when `enable_sql_audit` is set to `True`.
err_code	find_err_4013	Checks whether the error code `4013` is returned when `enable_sql_audit` is set to `True`.
err_code	find_err_4105	Checks whether the error code `4105` is returned when `enable_sql_audit` is set to `True`.
err_code	find_err_4012	Checks whether the error code `4012` is returned when `enable_sql_audit` is set to `True`.
err_code	find_err_4103	Checks whether the error code `4103` is returned when `enable_sql_audit` is set to `True`.
err_code	find_err_4001	Checks whether the error code `4001` is returned when `enable_sql_audit` is set to `True`.
err_code	find_err_4015	Checks whether the error code `4015` is returned when `enable_sql_audit` is set to `True`.
cpu	oversold	Checks whether CPU overprovisioning occurs on any OBServer nodes.
bugs	bug_182	After OceanBase Database is upgraded to V4.2.1, the error code `4109` indicating that the server status or role is not as expected is returned when a DDL statement is executed on some partitioned tables. This inspection metric is used to check for such scenarios. For more information, see issue #182 in GitHub.
bugs	bug_385	Checks for scenarios where a tenant has multiple root users when the version of OceanBase Database ranges from V4.2.1.0 to V4.2.1.3. In such scenarios, upgrade OceanBase Database to a later version or delete redundant users. For more information, see issue #385 in GitHub.
bugs	bug_469	Checks whether the ldd version is earlier than 2.34. For more information, see issue #469 in GitHub.
column_storage	tenant_parameters	Checks the tenant parameters that need to be optimized in the proof of concept (POC) scenario for columnar storage.
cluster	ls_nu	Checks whether the number of log stream replicas in the cluster is normal.
cluster	optimizer_better_inlist_costing_parmmeter	Checks whether the tag parameter of a specific version is enabled.
cluster	no_leader	Checks for leaderless log streams of any tenants in the cluster.
cluster	resource_limit_max_session_num	Check whether the `resourcelimit_max_sessionnum` parameter has been modified. If so, the error `Too many connections` may be returned.
cluster	deadlocks	Checks for deadlocks.
cluster	task_opt_stat	Checks the optimizer statistics collection history.
cluster	core_file_find	Checks for core dump files.
cluster	data_path_settings	Checks whether the data directory (specified by `data_dir`) and log directory (specified by `log_dir`) are on the same disk.
cluster	task_opt_stat_gather_fail	Checks for failed historical collection tasks.
cluster	part_trans_action_max	Checks whether the number of transaction participants exceeds 200.
cluster	zone_not_active	Checks whether the log stream (LS) ID is not_enough_replica.
cluster	table_history_too_many	Checks for tables that have excessive historical records. If tables of a tenant in the cluster have excessive historical records, error code `4013` is returned during architecture refresh when the server restarts. As a result, the corresponding tenant architecture cannot be refreshed on the server.
cluster	datafile_next	Checks the node parameter `datafile_maxsize`. If the `datafile_maxsize` parameter is set to a value greater than that of `datafile_size`, check whether the value of `datafile_next` is `0`. If so, the size of the data file will not grow. For more information, see issue #573 in GitHub.
cluster	major	Checks for suspended major compaction processes.
cluster	mod_too_large	Checks for modules that have used more than 10 GB of memory.
cluster	tenant_number	Checks the number of tenants.
cluster	global_indexes_too_much	Checks for tables that have more than 20 global indexes.
cluster	sys_log_level	Checks `sys_log_level`.
cluster	observer_not_active	Checks for OBServer nodes that are not in the `ACTIVE` state.
cluster	ob_enable_plan_cache_bad_version	Checks whether `ob_enable_plan_cache` is set to `0` to disable the plan cache in OceanBase Database V4.1.0.0 and V4.1.0.1 as recommended.
clog	clog_disk_full	Checks whether the clog disk is full.
sysbench	sysbench_test_tenant_log_disk_size	Checks the `log_disk_size` parameter of tenants.
sysbench	sysbench_run_test_tenant_cpu_used	Checks the CPU utilization of the cluster during a sysbench benchmark.
sysbench	sysbench_test_cluster_datafile_size	Checks the `datafile_size` and `log_disk_size` parameters of the cluster before a sysbench benchmark starts.
sysbench	sysbench_test_tenant_cpu_parameters	Checks the CPU parameters of tenants.
sysbench	sysbench_test_cluster_log_disk_size	Checks the `log_disk_size` parameter of the cluster.
sysbench	sysbench_run_test_tenant_memory_used	Checks the memory usage of tenants before a sysbench benchmark starts.
sysbench	sysbench_test_tenant_primary_zone	Checks the primary zone of the tenant during a sysbench benchmark.
sysbench	sysbench_test_sql_net_thread_count	Checks the `sql_net_thread_count` parameter of the cluster during a sysbench benchmark.
sysbench	sysbench_free_test_memory_limit	Checks the `memory_limit` parameter of the cluster before a sysbench benchmark starts.
sysbench	sysbench_free_test_network_speed	Checks the network speed of the cluster before a sysbench benchmark starts.
sysbench	sysbench_free_test_cpu_count	Checks the `cpu_count` parameter of the cluster before a sysbench benchmark starts.
sysbench	sysbench_test_log_level	Checks the `syslog_level` parameter of the cluster during a sysbench benchmark.
sysbench	sysbench_test_cluster_parameters	Checks the cluster parameters during a sysbench benchmark.
sysbench	sysbench_test_cpu_quota_concurrency	Checks the `cpu_quota_concurrency` parameter of the cluster during a sysbench benchmark.
network	network_drop	Checks the network packet loss information of the cluster.
network	TCP-retransmission	Checks the TCP retransmission information. For more information, see issue #348 in GitHub.
network	network_offset	Checks the network clock offset of the cluster.
network	network_speed	Checks the network speed of the cluster.
version	old_version	Checks the OceanBase Database version. Some OceanBase Database versions are not recommended.
version	bad_version	Checks the OceanBase Database version. Some OceanBase Database versions are not recommended.

ODP

Type	Task	Inspection purpose
version	bad_version	Checks whether an unrecommended OceanBase Database Proxy (ODP) version is used.
version	old_version	Checks whether an unrecommended ODP version is used.
parameter	request_buffer_length	Checks whether the default value of `request_buffer_length` is used.