Reliable message_等待事件分析 - 重庆思庄Oracle、Redhat认证学习论坛

db:11.2.0.4

现象：

Top 10 Foreground Events by Total Wait Time

Event	Waits	Total Wait Time (sec)	Wait Avg(ms)	% DB time	Wait Class
DB CPU		640.4		56.5
reliable message	193,457	179.2	1	15.8	Other
log file switch (checkpoint incomplete)	174	129.2	742	11.4	Configuration
gc current block 2-way	122,953	64.3	1	5.7	Cluster
enq: TX - index contention	7,662	35.6	5	3.1	Concurrency
gc buffer busy release	4,108	27.8	7	2.5	Cluster
gc current block busy	18,403	23.4	1	2.1	Cluster
buffer busy waits	5,378	22.6	4	2.0	Concurrency
SQL*Net more data from dblink	34,351	17.1	0	1.5	Network
gc buffer busy acquire	7,979	12	2	1.1	Cluster

具体可以参考《WAITEVENT: "reliable message" Reference Note (Doc ID 69088.1)》。

SQL> SELECT CHANNEL, SUM(wait_count) sum_wait_count

2 FROM GV$CHANNEL_WAITS

3 GROUP BY CHANNEL

4 ORDER BY SUM(wait_count) DESC;

CHANNEL SUM_WAIT_COUNT

---------------------------------------------------------------- --------------

Result Cache: Channel 1283422273

MMON remote action broadcast channel 1399425

kxfp control signal channel 1069929

RBR channel 925520

obj broadcast channel 507549

kill job broadcast - broadcast channel 960

parameters to cluster db instances - broadcast channel 960

LCK0 ksbxic channel 18

service operations - broadcast channel 2

quiesce channel 2

10 rows selected.

到此可以得出简单结论：该库的这个等待事件可能和Result Cache有关。

解决方案：

在mos上搜索相关关键字，发现了匹配的官方文档《Very High Waits for 'reliable message' After Upgrade to 11.2.0.4 When Using Result Cache (Doc ID 1951729.1)》，命中了bug 18416368。

当前版本是11.2.0.4版本，并配置了result cache，如下图：

SQL> show parameter result_cache_max_size

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

result_cache_max_size big integer 536096K

文档中给出了3种解决方案，若业务有影响，可根据情况选择其中一种方式进行修复：

升级到12C以上版本。

应用patch 18416368小补丁。

禁用result cache，并重启实例。

SQL> alter system set result_cache_max_size=0;