seabox inspect
seabox inspect¶
集群运维工具
- 语法
-
``` bash seabox inspect toolkit.
Usage: seabox inspect [OPTIONS]
Options: -o, –output cluster inspection report output file path -e, –expect inspect expect file -F, –format TEXT inspection report file format style -f, –full perform a complete inspect -S, –shrink pick files from log dir, shrink and move to log_shrink dir -n, –limit print the first K lines of the result -c, –check specified check item whose name is given in -l
-i –ignore ignore check item whose name is given in -i -v, –verbose output verbose message -l, –list list all check items
–start_time start_time start time for duration_query_check or shrink log, format by YYYY-mm-dd –end_time end_time end time for duration_query_check or shrink log, format by YYYY-mm-dd –write_lag write_lag write_lag for replication_lag_check, unit:second, format 0.000100 –flush_lag flush_lag flush_lag for replication_lag_check, unit:second, format 0.000100 –age_remain age_remain age_remain for database_age_check or relation_age_check, a integer value –time_consuming time_consuming time_consuming for active_xact_time_consuming_check, unit:second, a integer value –indexno indexno indexno for index_number_check, max number of index belong to one relation –relsize relsize relsize for index_number_check, unit:byte, max relation size –idxscan idxscan idxscan for index_use_check, times of idxscan, a integer value –idxsize idxsize idxsize for index_use_check, unit:byte, max index size –rollback_ratio rollback_ratio rollback_ratio for execute_stat_check, a float value –hit_ratio hit_ratio hit_ratio for execute_stat_check, a float value –conflicts conflicts conflicts for execute_stat_check, a integer value –deadlocks deadlocks deadlocks for execute_stat_check, a integer value –exact_size exact_size take longer to get an exact check –bloat_rate bloat_rate relation bloat rate –cv cv table size coefficient of variation –duration duration sql duration threshold, unit:seconds –audit_host audit_host sdaudit hostname –audit_port audit_port sdaudit port –audit_user audit_user sdaudit username –audit_pass audit_pass sdaudit user password –audit_dbname audit_dbname sdaudit database name -h, –help show this help, then exit
```
- 描述
-
集群运维工具集,包括集群巡检和审计日志处理等功能
- 注意:如有注意事项,补充
- 选项
-
该命令支持选项说明如下:
- Options
-
命令支持的选项
-o, --output TEXT
- 生成的巡检报告路径,如不指定则默认为
SDHOME/inspect_report
-e, --expect TEXT
- 预期文件路径
-F, --format TEXT
- 指定输出格式,默认为纯文本格式报告,可指定为json格式
-f, --full
- 执行完全巡检,包含耗时较长的检查
-S, --shrink
- 处理审计日志,提取重要日志并存放到log_shrink文件夹中
-n, --limit TEXT
- 输出结果的前n行
-c, --check TEXT
- 自定义巡检检查项,可指定多个检查项
-v, --verbose
- 详细的输出模式
-i --ignore
- 跳过某项检查
-l, --list
- 查看巡检中的所有检查项,用于-c
--start_time TIME --end_time TIME
- duration_query_check和shrink log检查项用来设置开始和结束时间,时间格式为YYYY-mm-dd,只在上述两项检查生效,可拆分指定
--write_lag LAG --flush_lag LAG
- 在replication_lag_check中指定write_lag和flush_lag的阈值,单位为秒,格式为数字,例如0.000100,只在上述检查生效,可拆分指定
--age_remain AGE
- 在database_age_check和relation_age_check用来设置age_remain的阈值,格式为整数值,只在上述两项检查生效
--time_consuming TIME
- 在active_xact_time_consuming_check用来设置time_consuming阈值,单位为秒,格式为整型,只在上述检查生效
--indexno NO --relsize SIZE
- 在index_number_check中用来设置indexno和relsize的阈值,格式为整型数值,relsize的单位为byte,只在上述检查生效,可拆分指定
--idxscan SCAN --idxsize SIZE
- 在index_use_check中用来设置idxscan和idxsize的阈值,格式为整型数值,idxsize单位为byte,只在上述检查生效,可拆分指定
--rollback_ratio R1 --hit_ratio R2 --conflicts C1 --deadlocks D1
- 在execute_stat_check用来指定以上四项指标的阈值,rollback_ratio和hit_ratio的格式为float类型,conflicts和deadlocks为整型数值,只在上述检查生效,可拆分指定
--exact_size VAL
- 指定relation_size_check的模式,当VAL为真时,会获取更准确的表容量,但是会花费更多的时间
--bloat_rate RATE
- 在relation_bloat_check中指定bloat_rate的阈值,格式为浮点型,只在上述检查生效
--cv VAL
- 在data_skew_check中设置变异系数的阈值,格式为浮点型,只在上述检查生效
--duration
- sql持续时间阀值,单位:秒
--audit_host
- sdaudit 主机名
--audit_port
- sdaudit 段口号
--audit_user
- sdaudit 用户名
--audit_pass
- sdaudit 用户密码
--audit_dbname
- sdaudit 数据库名称
- 示例
-
以下是
seabox inspect --help
的执行结果:seabox inspect toolkit. Usage: seabox inspect [OPTIONS] Options: -o, --output cluster inspection report output file path -e, --expect inspect expect file -F, --format TEXT inspection report file format style -f, --full perform a complete inspect -S, --shrink pick files from log dir, shrink and move to log_shrink dir -n, --limit print the first K lines of the result -c, --check specified check item whose name is given in -l -i, --ignore ignore check item whose name is given in -i -v, --verbose output verbose message -l, --list list all check items --start_time start_time start time for duration_query_check or shrink log, format by YYYY-mm-dd --end_time end_time end time for duration_query_check or shrink log, format by YYYY-mm-dd --write_lag write_lag write_lag for replication_lag_check, unit:second, format 0.000100 --flush_lag flush_lag flush_lag for replication_lag_check, unit:second, format 0.000100 --age_remain age_remain age_remain for database_age_check or relation_age_check, a integer value --time_consuming time_consuming time_consuming for active_xact_time_consuming_check, unit:second, a integer value --indexno indexno indexno for index_number_check, max number of index belong to one relation --relsize relsize relsize for index_number_check, unit:byte, max relation size --idxscan idxscan idxscan for index_use_check, times of idxscan, a integer value --idxsize idxsize idxsize for index_use_check, unit:byte, max index size --rollback_ratio rollback_ratio rollback_ratio for execute_stat_check, a float value --hit_ratio hit_ratio hit_ratio for execute_stat_check, a float value --conflicts conflicts conflicts for execute_stat_check, a integer value --deadlocks deadlocks deadlocks for execute_stat_check, a integer value --exact_size exact_size take longer to get an exact check --bloat_rate bloat_rate relation bloat rate --cv cv table size coefficient of variation --duration duration sql duration threshold, unit:seconds --audit_host audit_host sdaudit hostname --audit_port audit_port sdaudit port --audit_user audit_user sdaudit username --audit_pass audit_pass sdaudit user password --audit_dbname audit_dbname sdaudit database name -h, --help show this help, then exit
执行默认巡检,使用默认预期文件
seabox inspect -o report.out
执行默认巡检,指定预期文件和巡检报告路径
seabox inspect -e expect.yml -o report.out
执行完全巡检,指定预期文件和巡检报告路径
seabox inspect -e expect.yml -o report.out
执行默认巡检,指定输出json格式报告
seabox inspect -o report.out -F json
查看巡检中的所有检查项
seabox inspect -l base_info_check 数据库基本信息检查 key_dir_size_check 关键目录空间占用检查 etcd_space_check ETCD空间占用检查 key_config_diff_check 关键参数一致性检查 active_xact_time_consuming_check 活跃事务耗时检查 database_age_check 数据库年龄分析 cluster_connection_check 数据库连接分析 junk_data_check 数据库垃圾数据分析 execute_stat_check 数据库执行状态检查 replication_lag_check 节点流复制延迟检查 replication_state_check 节点流复制状态检查 bgwriter_stat_check 后台写状态检查 index_number_check 索引数量检查 index_use_check 索引使用检查 relation_age_check 数据库表年龄分析 relation_bloat_check 数据库表膨胀分析 duration_query_check 慢SQL检查 data_skew_check 数据倾斜检查 schema_space_check schema空间占用分析 database_space_check 数据库空间占用分析 tablespace_space_check 表空间空间占用分析 relation_size_check 数据库表容量分析 relation_index_exsits_check 无索引表分析 relation_primarykey_exsits_check 无主键表分析
自定义检查项执行巡检
seabox inspect -e expect.yml -o report.out -c base_info_check key_dir_size_check key_config_diff_check base_info_check ... normal (0.788s) key_dir_size_check ... normal (0.734s) key_config_diff_check ... normal (2.924s) seabox inspect passed
收缩n天前的审计日志
# 1. 配置巡检文件,指定 ndays_ago: n # 2. 执行收缩审计日志命令 seabox inspect -S
- 相关命令
- 无