rac的监控和调整

RAC的监控和调整

创建RAC的视图 $ORACLE_HOME/rdbms/admin/catclust.sql statspack报告里关于Cluster的统计： 1、GCS的工作负荷的特性 2、GES的统计信息 3、GCS和GES的信息统计 4、GES的详细统计 Select name,value from v$sysstat where name like ‘%global cache%’; Global Cache Service Request Latency 在一个完整Block请求中计算平台延时的公式（百分之一秒） Global Cache CR Block Receive Time Local Cache CR Blocks Received 1、Global Cache Service请求的延时，构成一个完整读块的平均时间： a = (global cache cr block build time) / (global cache cr blocks served) 2、等待Log Flush的平均时间 b = (global cache cr block flush time) / (global cache cr blocks served) 3、发送完整块的平均时间 c = (global cache cr block send time) / (global cache cr blocks served) 4、平均的lms服务时间 平均的延时 - a – b – c 5、监控当前的块的处理 当前的快的平均延时 = (global cache current block receive time / global cache current blocks received ) –( global cache current block pin time + global cache current block flush time + global cache current block send time ) / (global cache current blocks served) 如果值比较高，有可能在Buffers、cpu cycles、interconnect访问有可能竞争 6、监控block模式的转换 平均转换时间(毫秒) = 10 * (global cache convert time) / (global cache converts) 平均get时间(毫秒) = 10 * (global cache get time) / (global cache gets) 下面的统计计算了超时global cache convert timeouts，分析global enqueue统计 1. 收集global统计数据 a) 监控statspack ges统计部分 b) 查v$sysstat 2. 计算平均global enqueue时间 3. 计算平均global锁转换时间 4. 确定可能导致文件的资源类型 5. 使用v$librarycache 和v$rowcache视图来进行进一步的分析 select name, value from v$sysstat where name like ‘%global lock%’; 计算平均global enqueue时间（毫秒） 10 * (global lock get time) / (global lock sync gets + global lock async gets) 计算平均global lock convert时间（毫秒） 10 * (global lock convert time) / (global lock sync converts + global lock async converts) select event,time_waited,average_wait from v$system_event order by time_waited desc; select namespace, dlm_lock_request, dlm_pin,requests, dlm_pin_releasees, dlm_invalidation_requests, dlm_invalidations from v$librarycache; select parameter, dlm_requests, dlm_conflicts, dlm_releasees from v$rowcache; 监控global enqueue 服务资源统计 1、加入统计 EVENT = ” 29700 TRACE NAME CONTEXT FOREVER” 2、查统计结果 v$ges_convert_local v$ges_convert_remote 分析global enqueue服务资源统计 select r.convert_type, r.average_convert_time, l.average_convert_time, r.convert_count, l.convert_count from v$ges_convert_local l, v$ges_convert_remote r where r.convert_count <> 0 or l.convert_count <>0 group by r.convert_type; 其它的视图来查询资源活动 v$lock_activity v$class_cache_transfer v$cache_transfer v$file_cache_transfer V$system_evnet 事务的响应时间 response time/number of transaction = cpu time /number of transactions + wait time /number of transactions RAC在v$system_evnet中的事件 Global cache cr request Library cache pin Buffer busy due to global cache Global cache busy Global cache open x Global cache open s Global cache null to x Global cache s to x Global cache null to s 调整实例之间性能的观察 select paramater, count, dlm_requests, dlm_conflicts, dlm_releasees from v$rowcache; select namespace, dlm_lock_requests, dlm_pin_requests, dlm_pin_releases, dlm_invalidation_requests, dlm_invalidations from v$librarycache; 调优战略 收集基线数据 监控统计跟基线数据进行比较 只有在性能降低到不可以接受的水平才去调优 要监控跟cluster db无关的因素 address causes为更高水平的竞争