--- title: "分析 ASH 报告 - OceanBase 数据库 V4.3.0 | OceanBase 文档中心" description: 分析 ASH 报告 生成 ASH 报告后, 查看其内容以确定瞬时性能问题的可能原因。 ASH 报告内容 ASH 报告中通过 ASH Report 部分记录采样的起止时间、分析的起止时间、诊断的时间、样本的数量、事件的数量和平均活动会话数量这些信息，生成 Top User Events、Top Events P1/P2… --- 切换语言 - 中文站 - 简体中文 - International - English - 日本站 - 日本語 文档反馈![](https://mdn.alipayobjects.com/huamei_22khvb/afts/img/A*P8CuR4UJ_FkAAAAAAAAAAAAADiGDAQ/original) OceanBase 数据库分布式版 - V 4.3.0 # 分析 ASH 报告 更新时间：2026-04-10 11:58:35 [编辑](https://github.com/oceanbase/oceanbase-doc/edit/V4.3.0/zh-CN/700.reference/1000.performance-tuning-guide/400.performance-diagnosis/500.ash-report-diagnosis/300.analyze-ash-report.md) 生成 ASH 报告后, 查看其内容以确定瞬时性能问题的可能原因。 ## ASH 报告内容 ASH 报告中通过 ASH Report 部分记录采样的起止时间、分析的起止时间、诊断的时间、样本的数量、事件的数量和平均活动会话数量这些信息，生成 Top User Events、Top Events P1/P2/P3 Value、Top Phase of Execution、Top SQL with Top Events、Top SQL with Top Blocking Events、Complete List of SQL Text、Top Sessions、Top Blocking Sessions、Top latches 等 9 个模块的信息，从多个维度或组合（例如时间、会话、模块、操作或 SQL 语句）进行范围或目标的性能分析。 ```shell # ASH Report ---- Sample Begin: 2024-01-05 21:19:00 Sample End: 2024-01-09 13:58:02 ---------- Analysis Begin Time: 2024-01-06 11:17:25 Analysis End Time: 2024-01-09 13:57:49 Elapsed Time: 268824(secs) Num of Sample: 10814 Num of Events: 10814 Average Active Sessions: 0.040 ---- ## Top User Events: ----------------------------------------------------------------+--------------------+---------+ Event| WAIT_CLASS| % Event| ----------------------------------------------------------------+--------------------+---------+ CPU + Wait for CPU| OTHER| 100.00%| ----------------------------------------------------------------+--------------------+---------+ ## Top Events P1/P2/P3 Value: ----------------------------------------------------------------+----------+------------+--------------------------------------------------+--------------------+--------------------+--------------------+ Event| % Event| % Activity| Max P1/P2/P3| Parameter 1| Parameter 2| Parameter 3| ----------------------------------------------------------------+----------+------------+--------------------------------------------------+--------------------+--------------------+--------------------+ CPU + Wait for CPU| 100.00%| 100.000%| "0","0","0"| | | | ----------------------------------------------------------------+----------+------------+--------------------------------------------------+--------------------+--------------------+--------------------+ ## Top Phase of Execution: ----------------------------------------+------------+--------------+----------------------------------------+ Phase of Execution| % Activity| Sample Count| Avg Active Sessions| ----------------------------------------+------------+--------------+----------------------------------------+ IN_SQL_EXECUTION| 0.203%| 22| 0.00| IN_SQL_OPTIMIZE| 0.009%| 1| 0.00| IN_PARSE| 0.000%| 0| 0.00| IN_PL_PARSE| 0.000%| 0| 0.00| IN_PLAN_CACHE| 0.000%| 0| 0.00| IN_PX_EXECUTION| 0.000%| 0| 0.00| IN_SEQUENCE_LOAD| 0.000%| 0| 0.00| IN_COMMITTING| 0.000%| 0| 0.00| IN_STORAGE_READ| 0.000%| 0| 0.00| IN_STORAGE_WRITE| 0.000%| 0| 0.00| IN_REMOTE_DAS_EXECUTION| 0.000%| 0| 0.00| ----------------------------------------+------------+--------------+----------------------------------------+ ## Top SQL with Top Events - All events included. - Empty 'SQL Text' if it is PL/SQL query ----------------------------------------+------------+-------------------------+----------------------------------------------------------------+------------+------------------------------------------------------------+ SQL ID| PLAN ID| Sampled # of Executions| Event| % Event| SQL Text| ----------------------------------------+------------+-------------------------+----------------------------------------------------------------+------------+------------------------------------------------------------+ | 0| 10813| CPU + Wait for CPU| 99.99%| | 15B18561C66FCF844D00386749C0DA49| 1932| 1| CPU + Wait for CPU| 0.01%| | ----------------------------------------+------------+-------------------------+----------------------------------------------------------------+------------+------------------------------------------------------------+ ## Top SQL with Top Blocking Events - Empty result if no event other than On CPU sampled - Empty 'SQL Text' if it is PL/SQL query ----------------------------------------+------------+-------------------------+----------------------------------------------------------------+------------+------------------------------------------------------------+ SQL ID| PLAN ID| Sampled # of Executions| Event| % Event| SQL Text| ----------------------------------------+------------+-------------------------+----------------------------------------------------------------+------------+------------------------------------------------------------+ ----------------------------------------+------------+-------------------------+----------------------------------------------------------------+------------+------------------------------------------------------------+ ## Complete List of SQL Text ## Top PL/SQL Procedures -> 'PL/SQL entry subprogram' represents the application's top-level entry-point(procedure, function, trigger, package initialization) into PL/SQL. -> 'PL/SQL current subprogram' is the pl/sql subprogram being executed at the point of sampling . If the value is 'SQL', it represents the percentage of time spent executing SQL for the particular plsql entry subprogram -> 'PL/SQL entry subprogram' represents the application's top-level ------------------------------------------------------------+------------------------------------------------------------+--------------------+ PLSQL Entry Subprogram| PLSQL Current Subprogram| % Activity| ------------------------------------------------------------+------------------------------------------------------------+--------------------+ ------------------------------------------------------------+------------------------------------------------------------+--------------------+ ## Top Sessions: - '# Samples Active' shows the number of ASH samples in which the session was found waiting for that particular event. The percentage shown in this column is calculated with respect to wall time. --------------------+----------------------+----------------------------------------------------------------+------------+------------+--------------------+--------------------+ Sid| % Activity| Event| Event Count| % Event| User| # Samples Active| --------------------+----------------------+----------------------------------------------------------------+------------+------------+--------------------+--------------------+ 4436701216768| 48.71%| CPU + Wait for CPU| 5267| 48.71%| | 5267/268824[1.96%]| 4398046511104| 5.84%| CPU + Wait for CPU| 632| 5.84%| | 632/268824[0.24%]| 3861175599104| 5.04%| CPU + Wait for CPU| 545| 5.04%| | 545/268824[0.20%]| 5312874545152| 5.02%| CPU + Wait for CPU| 543| 5.02%| | 543/268824[0.20%]| 5317169512448| 5.02%| CPU + Wait for CPU| 543| 5.02%| | 543/268824[0.20%]| 3865470566400| 4.29%| CPU + Wait for CPU| 464| 4.29%| | 464/268824[0.17%]| 4471060955136| 2.26%| CPU + Wait for CPU| 244| 2.26%| | 244/268824[0.09%]| 3925600108544| 1.81%| CPU + Wait for CPU| 196| 1.81%| | 196/268824[0.07%]| 3981434683392| 1.54%| CPU + Wait for CPU| 167| 1.54%| | 167/268824[0.06%]| 4264902524928| 1.36%| CPU + Wait for CPU| 147| 1.36%| | 147/268824[0.05%]| 4320737099776| 1.28%| CPU + Wait for CPU| 138| 1.28%| | 138/268824[0.05%]| 4458176053248| 1.04%| CPU + Wait for CPU| 112| 1.04%| | 112/268824[0.04%]| 4402341478400| 0.97%| CPU + Wait for CPU| 105| 0.97%| | 105/268824[0.04%]| 4256312590336| 0.92%| CPU + Wait for CPU| 100| 0.92%| | 100/268824[0.04%]| 3813930958848| 0.91%| CPU + Wait for CPU| 98| 0.91%| | 98/268824[0.04%]| 3856880631808| 0.83%| CPU + Wait for CPU| 90| 0.83%| | 90/268824[0.03%]| 3934190043136| 0.80%| CPU + Wait for CPU| 87| 0.80%| | 87/268824[0.03%]| 3938485010432| 0.80%| CPU + Wait for CPU| 86| 0.80%| | 86/268824[0.03%]| 4350801870848| 0.78%| CPU + Wait for CPU| 84| 0.78%| | 84/268824[0.03%]| --------------------+----------------------+----------------------------------------------------------------+------------+------------+--------------------+--------------------+ ## Top Blocking Sessions: - Blocking session activity percentages are calculated with respect to waits on latches and locks only. - '# Samples Active' shows the number of ASH samples in which the blocking session was found active. --------------------+----------------------+----------------------------------------+------------+------------+--------------------+--------------------+ Blocking Sid| % Activity| Event Caused| Event Count| % Event| User| # Samples Active| --------------------+----------------------+----------------------------------------+------------+------------+--------------------+--------------------+ --------------------+----------------------+----------------------------------------+------------+------------+--------------------+--------------------+ ## Top latches: ----------------------------------------------------------------+--------------------+--------------------+ Latch| Sampled Count| % Activity| ----------------------------------------------------------------+--------------------+--------------------+ ----------------------------------------------------------------+--------------------+--------------------+ ## Node Load: ----------------------------------------+--------------------+--------------------+--------------------+------------------------------+ SVR IP| SVR PORT| Sampled Count| Idle Wait Count| Load| ----------------------------------------+--------------------+--------------------+--------------------+------------------------------+ xx.xx.xx.xx| 28824| 10814| 0| 0.04| ----------------------------------------+--------------------+--------------------+--------------------+------------------------------+ ``` ### Top User Events 这部分内容通过事件、事件类型、采样事件个数和等待事件占总时间的百分比等多个维度，罗列出了占采样会话活动百分比最高的用户进程中排名靠前的等待事件。 #### 说明 - CPU + Wait for CPU ： CPU 的正常执行情况。 - sync rpc ：同步等待 RPC 网络包。 ### Top Events P1/P2/P3 Value 这部分显示了在采样会话活动中占比最大的等待事件的参数值, 与总的等待时间 （%Event）百分比的排序。对于每一个等待事件 p1，p2，p3 的值对应 parameter 1，parameter 2，parameter 3 列中显示的等待事件参数。 ### Top Phase of Execution 这部分显示了对采样会话中执行阶段、活动会话百分比、 样本计数、和平均活动会话数之间的比较，得到活动百分比最高的执行阶段的排序，例如 SQL、PL/SQL 和 Java 编译和执行的过程。 ### Top SQL with Top Events 这部分显示占采样会话活动百分比最高的 SQL 语句以及这些 SQL 语句遇到的最常见等待事件。Sampled # of Executions 列指对特定 SQL 语句的不同执行次数进行了采样。 ### Top SQL with Top Blocking Events 这部分显示被阻塞的会话信息。可以根据这块信息确定 SQL 执行的哪一部分对 SQL 运行时间有显着影响。 ### Complete List of SQL Text 这部分显示 Top SQL 中 SQL 语句的完整文本内容。 ### Top Sessions 这部分以 Session 为维度进行排序统计活跃的 Session。通过采样的会话 ID、活动会话百分比、事件、事件个数、事件百分比，当前用户和采样的活动状态之间的比较，方便找到潜藏的问题 Session。 ### Top Blocking Sessions 这部分显示占采样会话活动百分比最高的阻止会话。 ### Top latches 这部分显示了在采样会话活动中百分比最高的闩锁。 #### 说明 闩锁是简单的低级序列化机制，用于保护系统全局区中的共享数据结构。 例如，锁存器保护当前访问数据库的用户列表和描述缓冲区缓存中块的数据结构。服务器或后台进程在操作或查看这些结构之一时会在很短的时间内获取闩锁。锁存器的实现依赖于操作系统，特别是关于进程是否等待锁存器以及等待多长时间。 ## 常见的等待事件 一些常见的等待事件总结如下，方便您在分析报告时，快速的定位问题。 | **等待事件名称** | **描述** | **等待事件所属类型名称** | **等待事件的参数 1** | **等待事件的参数 2** | **等待事件的参数 3** | | --- | --- | --- | --- | --- | --- | | wait remove partition | 分区 GC 等待事务模块删除 | ADMINISTRATIVE | tenant_id | table_id | partition_id | | wait end stmt | 语句结束等待 | CLUSTER | rollback | trans_hash_value | physic_plan_type | | wait start stmt | 语句开启等待 | CLUSTER | trans_hash_value | physic_plan_type | participant_count | | wait end trans | 事务提交等待 | COMMIT | rollback | trans_hash_value | participant_count | | async rpc proxy condition wait | 上层封装的异步 RPC Wait 事件 | CONCURRENCY | address | - | - | | dag worker condition wait | 后台 dag 线程等待事件 | CONCURRENCY | address | - | - | | hashmap lock wait | HASH 表锁等待事件 | CONCURRENCY | address | number | tries | | latch: alive server tracer lock wait | OBAliveServerMap 操作内存中 OBServer 状态的锁等待 | CONCURRENCY | address | number | tries | | latch: alloc block lock wait | OceanBase malloc 内部管理 K 级别内存块的锁等待 | CONCURRENCY | address | number | tries | | latch: alloc object lock wait | OceanBase malloc 内部管理小块内存的锁等待事件 | CONCURRENCY | address | number | tries | | latch: allocator chunk lock wait | OceanBase malloc 内部管理 2 M 以上内存的锁等待事件 | CONCURRENCY | address | number | tries | | latch: clog cascading info lock wait | clog 副本级联模块锁的等待事件 | CONCURRENCY | address | number | tries | | latch: clog locality lock wait | clog 内部保护 Locality 信息的锁 | CONCURRENCY | address | number | tries | | latch: clog membership mgr lock wait | clog 模块成员变更的锁 | CONCURRENCY | address | number | tries | | latch: clog reconfirm lock wait | clog 模块 Reconfirm 类使用的锁 | CONCURRENCY | address | number | tries | | latch: clog stat mgr lock wait | clog 模块状态切换的锁 | CONCURRENCY | address | number | tries | | latch: clog sw renew ms task lock wait | clog 备库更新成员组信息的锁等待 | CONCURRENCY | address | number | tries | | latch: clog switch info lock wait | clog 模块保护物理备库中主备租户切换信息的锁 | CONCURRENCY | address | number | tries | | latch: clog task lock wait | clog 模块 log_task 的锁 | CONCURRENCY | address | number | tries | | latch: config lock wait | 配置项锁等待事件 | CONCURRENCY | address | number | tries | | latch: default drw lock wait | DRWLock 等待事件，DRWLock 是一个适应读多写少场景的读写锁 | CONCURRENCY | address | number | tries | | latch: default mutex wait | 缺省 MUTEX 等待事件 | CONCURRENCY | address | number | tries | | latch: default recursive mutex wait | 缺省递归 MUTEX 等待事件 | CONCURRENCY | address | number | tries | | latch: default spin lock wait | 缺省自旋锁等待事件 | CONCURRENCY | address | number | tries | | latch: default spin rwlock wait | 缺省自旋读写锁等待事件 | CONCURRENCY | address | number | tries | | latch: id map node lock wait | ID Map 等待事件，ID Map 用于把整数映射为指针 | CONCURRENCY | address | number | tries | | latch: plan cache pcv set lock wait | Plan Cache 添加新对象时加锁等待耗时 | CONCURRENCY | address | number | tries | | latch: plan set lock wait | Plan Cache 加锁耗时 | CONCURRENCY | address | number | tries | | latch: ps store lock wait | PS Cache 加锁耗时 | CONCURRENCY | address | number | tries | | latch: server maintaince lock wait | RS 进行 OBServer 操作时的锁等待 | CONCURRENCY | address | number | tries | | latch: server status lock wait | 操作 RS 记录的 server_statuses 时的读写锁等待 | CONCURRENCY | address | number | tries | | latch: time wheel bucket lock wait | 事务层定时器上 BUCKET 锁 | CONCURRENCY | address | number | tries | | latch: time wheel task lock wait | 事务定时任务的锁 | CONCURRENCY | address | number | tries | | latch: trans ctx lock wait | 事务上下文并发操作的互斥锁等待 | CONCURRENCY | address | number | tries | | latch: unit manager lock wait | Unit Manager 操作数据的锁等待 | CONCURRENCY | address | number | tries | | latch: zone infos rw lock wait | OBZoneManagerBase 操作数据的锁等待 | CONCURRENCY | address | number | tries | | latch: zone manager maintaince lock wait | Zone Manager 操作数据的锁等待 | CONCURRENCY | address | number | tries | | ob dynamic thread pool condition wait | 动态扩展的线程池锁等待事件 | CONCURRENCY | address | - | - | | partition location cache lock wait | Location Cache 刷新时，防并发刷新等待事件 | CONCURRENCY | table_id | partition_id | - | | partition table updater condition wait | OBUniqTaskQueue 处理任务时的条件等待 | CONCURRENCY | address | - | - | | rebalance task mgr condition wait | OBDRTaskMgr 处理容灾任务时的条件等待 | CONCURRENCY | address | - | - | | reentrant thread condition wait | OBReentrantThread 处理任务时的条件等待 | CONCURRENCY | address | - | - | | rpc session handler condition wait | 流式 RPC 等待事件 | CONCURRENCY | address | - | - | | th worker condition wait | 多租户线程挂起等待事件 | CONCURRENCY | address | - | - | | thread idling condition wait | RootService 空闲线程等待事件 | CONCURRENCY | address | - | - | | sched idle | 多租户线程等待任务执行 | IDLE | wait start timestamp | - | - | | mysql response wait client | 等待客户端返回耗时 | NETWORK | - | - | - | | sync rpc | 同步 RPC 等待事件 | NETWORK | pcode | size | - | | sched wait | 多租户线程因执行大查询被挂起 | SCHEDULER | req type | req start timestamp | wait start timestamp | | bloomfilter build read | 后台 BloomFilter 创建数据读取 | SYSTEM_IO | fd | offset | size | | db file compact read | compaction 任务数据读取 | SYSTEM_IO | fd | offset | size | | db file compact write | compaction 任务数据写入 | SYSTEM_IO | fd | offset | size | | db file index build read | 索引创建数据读取 | SYSTEM_IO | fd | offset | size | | db file index build write | 索引创建数据写入 | SYSTEM_IO | fd | offset | size | | db file migrate read | 迁移数据读取 | SYSTEM_IO | fd | offset | size | | db file migrate write | 迁移数据写入 | SYSTEM_IO | fd | offset | size | | memstore memory page alloc info | MemStore 写入限速触发的计算值 | SYSTEM_IO | cur_mem_hold | sleep_interval | cur_ts | | memstore memory page alloc wait | MemStore 写入限速触发的等待时间 | SYSTEM_IO | cur_mem_hold | sleep_interval | cur_ts | | db file data index read | 用户数据微块索引读取 | USER_IO | fd | offset | size | | db file data read | 用户数据读取 | USER_IO | fd | offset | size | | interm result disk read | 中间结果数据读取 | USER_IO | fd | offset | size | | interm result disk write | 中间结果数据写入 | USER_IO | fd | offset | size | | row store disk read | SQL 算子落盘数据读取 | USER_IO | fd | offset | size | | row store disk write | SQL 算子落盘数据写入 | USER_IO | fd | offset | size | | backup info wait | 备份信息管理器内部的等待 | CONCURRENCY | address | number | tries | | build index scheduler condition wait | DDL 两次任务调度的等待时间 | CONCURRENCY | address | - | - | | debug sync condition wait | Errsim Debug 模式下有意义，用户场景无意义 | CONCURRENCY | address | - | - | | inner connection pool condition wait | 内部连接池条件等待 | CONCURRENCY | address | - | - | | latch: bandwidth throttle lock wait | 带宽限流的锁等待 | CONCURRENCY | address | number | tries | | latch: default bucket lock wait | Bucket Lock 的等待时间，未归类 | CONCURRENCY | address | number | tries | | latch: election group lock wait | 保护 Election Group 信息的锁 | CONCURRENCY | address | number | tries | | latch: election group trace recorder lock wait | 保护 Election Group Trace 统计信息的锁 | CONCURRENCY | address | number | tries | | latch: election lock wait | 选主线程的等锁时间 | CONCURRENCY | address | number | tries | | latch: election msg lock wait | 选举消息的锁等待 | CONCURRENCY | address | number | tries | | latch: group migrate lock wait | 批量迁移的锁等待 | CONCURRENCY | address | number | tries | | latch: group migrate task idle wait | 批量迁移的任务空闲锁等待 | CONCURRENCY | address | number | tries | | latch: group migrate task lock wait | 批量迁移的任务锁等待 | CONCURRENCY | address | number | tries | | latch: io queue lock wait | I/O queue 读取时的锁等待 | CONCURRENCY | address | number | tries | | latch: migrate lock wait | 迁移的锁等待 | CONCURRENCY | address | number | tries | | latch: multi tenant lock wait | 多租户管理的锁等待 | CONCURRENCY | address | number | tries | | latch: rebuild retry list lock wait | Rebuild 任务的等待 | CONCURRENCY | address | number | tries | | latch: session trace recorder lock wait | Session 上 Trace 记录的锁等待 | CONCURRENCY | address | number | tries | | latch: tenant lock wait | 租户内管理的锁等待 | CONCURRENCY | address | number | tries | | latch: timezone lock wait | 租户的 Timezone 结构锁等待，租户操作的时候会上锁 | CONCURRENCY | address | number | tries | | latch: trace recorder lock wait | SQL Trace 记录的锁等待； | CONCURRENCY | address | number | tries | | ob request timeinfo list wait | 请求计数的锁统计，跟 Plan Cache 兜底有关 | CONCURRENCY | address | - | - | | ob restore reader condition wait | 备份恢复相关的等待信息 | CONCURRENCY | address | - | - | | oblog part mgr schema version wait | Liboblog 的等待刷新 Schema Verison 的逻辑 | CONCURRENCY | - | - | - | | slog flush condition wait | Slog 刷盘的等待时间 | CONCURRENCY | address | - | - | | ObCacheLineSegregatedArrayBase alloc mutex | Cache Line 数组分配锁等待 | USER_IO | fd | offset | size | | server object pool arena lock wait | 大对象缓存结构的锁等待 | USER_IO | fd | offset | size | ## 相关文档 - [ASH 简介](https://www.oceanbase.com/docs/common-oceanbase-database-cn-1000000000640561)。 - [生成 ASH 报告](https://www.oceanbase.com/docs/common-oceanbase-database-cn-1000000000640563)。 上一篇 下一篇 ![有帮助](https://gw.alipayobjects.com/mdn/ob_asset/afts/img/A*y6ocSqN8cqsAAAAAAAAAAAAAARQnAQ)![无帮助](https://gw.alipayobjects.com/mdn/ob_asset/afts/img/A*BG9IQJyLHF8AAAAAAAAAAAAAARQnAQ)![反馈](https://gw.alipayobjects.com/mdn/ob_asset/afts/img/A*eTWdQKCRKHwAAAAAAAAAAAAAARQnAQ)[AI](https://www.oceanbase.com/obi) 咨询热线