本文永久地址https://www.askmac.cn/archives/rac-awr-statistics.html

RAC相关指标

Global Cache Load Profile

	Per Second	Per Transaction
Global Cache blocks received:	12.06	2.23
Global Cache blocks served:	8.18	1.51
GCS/GES messages received:	391.19	72.37
GCS/GES messages sent:	368.76	68.22
DBWR Fusion writes:	0.10	0.02
Estd Interconnect traffic (KB)	310.31

指标	指标说明
Global Cache blocks received	通过硬件连接收到远程实例的数据块的数量。发生在一个进程请求一致性读一个数据块不是在本地缓存中。Oracle发送一个请求到另外的实例。一旦缓冲区收到，这个统计值就会增加。这个统计值是另两个统计值的和：Global Cache blocks received = gc current blocks received + gc cr blocks received
Global Cache blocks served	通过硬件连接发送到远程实例的数据块的数量。这个统计值是另外两个统计值的和：Global Cache blocks served = gc current blocks served + gc cr blocks served
GCS/GES messages received	通过硬件连接收到远程实例的消息的数量。这个统计值通常代表RAC服务引起的开销。这个统计值是另外两个统计值的和：GCS/GES messages received = gcs msgs received + ges msgs received
GCS/GES messages sent	通过硬件连接发送到远程实例的消息的数量。这个统计值通常代表RAC服务引起的开销。这个统计值是另外两个统计值的和：GCS/GES messages sent = gcs messages sent + ges messages sent
DBWR Fusion writes	这个统计值显示融合写入的次数。在RAC中，单实例Oracle数据库，数据块只被写入磁盘因为数据过期，缓冲替换或者发生检查点。当一个数据块在缓存中被替换因为数据过期或发生检查点但在另外的实例没有写入磁盘，Global Cache Service会请求实例将数据块写入磁盘。因此融合写入不包括在第一个实例中的额外写入磁盘。大量的融合写入表明一个持续的问题。实例产生的融合写入请求占总的写入请求的比率用于性能分析。高比率表明DB cache大小不合适或者检查点效率低。
Estd Interconnect traffic (KB)	连接传输的KB大小。计算公式如下：Estd Interconnect traffic (KB) = ((‘gc cr blocks received’+ ‘gc current blocks received’ + ‘gc cr blocks served’+ ‘gc current blocks served’) * Block size) + ((‘gcs messages sent’ + ‘ges messages sent’ + ‘gcs msgs received’+ ‘gcs msgs received’)*200)/1024/Elapsed Time

Global Cache Efficiency Percentages (Target local+remote 100%)

Buffer access – local cache %:	91.05
Buffer access – remote cache %:	0.03
Buffer access – disk %:	8.92

指标	指标说明
Buffer access – local cache %	数据块从本地缓存命中占会话总的数据库请求次数的比例。在OLTP应用中最希望的是尽可能维持这个比率较高，因为这是最低成本和最快速的获得数据库数据块的方法。计算公式：Local Cache Buffer Access Ratio = 1 – ( physical reads cache + Global Cache blocks received ) / Logical Reads
Buffer access – remote cache %	数据块从远程实例缓存命中占会话总的数据块请求的比例。在OLTP应用中这个比率和Buffer access – local cache的和应该尽可能的高因为这两种方法访问数据库数据块是最快速最低成本的。这个比率的计算方法：Remote Cache Buffer Access Ratio = Global Cache blocks received / Logical Reads
Buffer access – disk %	从磁盘上读数据块到缓存占会话总的数据块请求次数的比例。在OLTP应用中希望维持这个比例低因为物理读是最慢的访问数据库数据块的方式。这个比率计算方法： 1 – physical reads cache / Logical Reads

Global Cache and Enqueue Services – Workload Characteristics

Avg global enqueue get time (ms):	0.0
Avg global cache cr block receive time (ms):	0.3
Avg global cache current block receive time (ms):	0.2
Avg global cache cr block build time (ms):	0.0
Avg global cache cr block send time (ms):	0.0
Global cache log flushes for cr blocks served %:	1.2
Avg global cache cr block flush time (ms):	1.8
Avg global cache current block pin time (ms):	1,021.7
Avg global cache current block send time (ms):	0.0
Global cache log flushes for current blocks served %:	6.9
Avg global cache current block flush time (ms):	0.9

本文永久地址https://www.askmac.cn/archives/rac-awr-statistics.html

指标	指标说明
Avg global enqueue get time (ms)	通过interconnect发送消息，为争夺资源开启一个新的全局队列或者对已经开启的队列转换访问模式所花费的时间。如果大于20ms，你的系统可能会出现超时。
Avg global cache cr block receive time (ms)	从请求实例发送消息到mastering instance（2-way get）和一些到holding instance (3-way get)花费的时间。这个时间包括在holding instance生成数据块一致性读映像的时间。CR数据块获取耗费的时间不应该大于15ms。
Avg global cache current block receive time (ms)	从请求实例发送消息到mastering instance（2-way get）和一些到holding instance (3-way get)花费的时间。这个时间包括holding instance日志刷新花费的时间。Current Block获取耗费的时间不大于30ms
Avg global cache cr block build time (ms)	CR数据块创建耗费的时间
Avg global cache cr block send time (ms)	CR数据块发送耗费的时间
Global cache log flushes for cr blocks served %	需要日志刷新的CR数据块占总的需要服务的CR数据块的比例。
Avg global cache cr block flush time (ms)	CR数据块刷新耗费的时间
Avg global cache current block pin time (ms)	Current数据块pin耗费的时间
Avg global cache current block send time (ms)	Current数据块发送耗费的时间
Global cache log flushes for current blocks served %	需要日志刷新的Current数据块占总的需要服务的Current数据块的比例
Avg global cache current block flush time (ms)	Current数据块刷新耗费的时间

Global Cache and Enqueue Services – Messaging Statistics

Avg message sent queue time (ms):	2,367.6
Avg message sent queue time on ksxp (ms):	0.1
Avg message received queue time (ms):	0.3
Avg GCS message process time (ms):	0.0
Avg GES message process time (ms):	0.0
% of direct sent messages:	54.00
% of indirect sent messages:	44.96
% of flow controlled messages:	1.03

指标	指标说明
Avg message sent queue time (ms)	一条信息进入队列到发送它的时间
Avg message sent queue time on ksxp (ms)	对端收到该信息并返回ACK的时间，这个指标很重要，直接反应了网络延迟，一般小于1ms
Avg message received queue time (ms)	一条信息进入队列到收到它的时间
Avg GCS message process time (ms)
Avg GES message process time (ms)
% of direct sent messages	直接发送信息占的比率
% of indirect sent messages	间接发送信息占的比率，一般是排序或大的信息，流控制也可能引起
% of flow controlled messages	流控制信息占的比率，流控制最常见的原因是网络状况不佳， % of flow controlled messages应当小于1%

Wait Event Histogram

	% of Waits
Event	Total Waits	<1ms	<2ms	<4ms	<8ms	<16ms	<32ms	<=1s	>1s
ADR block file read	208	38.0		3.4	44.7	13.9
ADR block file write	40	100.0
ADR file lock	48	100.0
ARCH wait for archivelog lock	3	100.0
ASM file metadata operation	12.8K	99.7	.1	.0		.0	.0	.2	.0
Backup: MML write backup piece	310.5K	7.6	.1	.1	1.3	10.4	30.2	50.2	.0
CGS wait for IPC msg	141.7K	100.0
CSS initialization	34	50.0			47.1	2.9
CSS operation: action	110	48.2	20.9	28.2	2.7
CSS operation: query	102	88.2	3.9	7.8
DFS lock handle	6607	93.9	.5	.2	.0		.0	5.3	.0
Disk file operations I/O	1474	100.0
IPC send completion sync	21.9K	99.5	.1	.1	.1	.0	.2
KJC: Wait for msg sends to complete	13	100.0
LGWR wait for redo copy	16.3K	100.0	.0
Log archive I/O	3	33.3	66.7
PX Deq: Signal ACK EXT	2256	99.8	.1	.1
PX Deq: Signal ACK RSG	2124	99.9	.1	.0
PX Deq: Slave Session Stats	7997	94.6	.9	.9	2.5	.8	.4
PX Deq: Table Q qref	2355	99.9	.1
PX Deq: reap credit	1215.7K	100.0	.0	.0
PX qref latch	1366	100.0
Parameter File I/O	194	94.8	1.0		1.0	1.0		1.5	.5

Wait Event Histogram：等待时间直方图

Event：等待事件名字

Total Waits：该等待事件在快照时间内等待的次数

%of Waits < 1ms ：小于1ms的等待次数

%of Waits < 2ms ：小于2ms的等待次数

%of Waits < 4ms ：小于4ms的等待次数

%of Waits < 8ms ：小于8ms的等待次数

%of Waits < 16ms ：小于16ms的等待次数

%of Waits < 32ms ：小于32ms的等待次数

%of Waits < =1s ：小于等于1s的等待次数

%of Waits > 1s ：大于1s的等待次数

Parent Latch Statistics

only latches with sleeps are shown
ordered by name

Latch Name	Get Requests	Misses	Sleeps	Spin & Sleeps 1->3+
Real-time plan statistics latch	77,840	136	20	116/0/0/0
active checkpoint queue latch	321,023	20,528	77	20451/0/0/0
active service list	339,641	546	132	424/0/0/0
call allocation	328,283	550	148	440/0/0/0
enqueues	1,503,525	217	14	203/0/0/0
ksuosstats global area	2,605	1	1	0/0/0/0
messages	2,608,863	141,380	29	141351/0/0/0
name-service request queue	155,047	43	15	28/0/0/0
qmn task queue latch	2,368	90	78	12/0/0/0
query server process	268	30	30	0/0/0/0
redo writing	910,703	11,623	50	11573/0/0/0
resmgr:free threads list	14,454	190	4	186/0/0/0
space background task latch	11,209	15	7	8/0/0/0

Latch Name：闩名称

Get Requests：申请获得父闩的次数

本文永久地址https://www.askmac.cn/archives/rac-awr-statistics.html

Child Latch Statistics

only latches with sleeps/gets > 1/100000 are shown
ordered by name, gets desc

Latch Name	Child Num	Get Requests	Misses	Sleeps	Spin & Sleeps 1->3+
KJC message pool free list	1	96,136	82	20	62/0/0/0
Lsod array latch	10	2,222	153	118	58/0/0/0
Lsod array latch	13	2,151	43	14	29/0/0/0
Lsod array latch	4	2,066	154	124	59/0/0/0
Lsod array latch	5	1,988	105	44	63/0/0/0
Lsod array latch	9	1,734	95	32	64/0/0/0
Lsod array latch	2	1,707	88	38	55/0/0/0
Lsod array latch	11	1,695	88	32	57/0/0/0
Lsod array latch	6	1,680	158	126	64/0/0/0
Lsod array latch	12	1,657	155	111	65/0/0/0
Lsod array latch	7	1,640	90	34	59/0/0/0
Lsod array latch	1	1,627	169	153	46/0/0/0
Lsod array latch	3	1,555	87	36	54/0/0/0
Lsod array latch	8	1,487	127	88	57/0/0/0
cache buffers chains	47418	354,313	391	4	387/0/0/0
cache buffers chains	8031	337,135	250	8	242/0/0/0
cache buffers chains	78358	305,022	528	9	519/0/0/0
cache buffers chains	6927	241,808	129	4	125/0/0/0

Latch Name：闩名称

Child Num：

Get Requests：

Misses：

Sleeps：

Spin&Sleeps 1->3+：

Dictionary Cache Stats (RAC)

Cache	GES Requests	GES Conflicts	GES Releases
dc_awr_control	11	5	0
dc_global_oids	5	0	0
dc_histogram_defs	215	1	707
dc_objects	90	9	0
dc_segments	79	10	73
dc_sequences	35,738	37	0
dc_table_scns	6	0	0
dc_tablespace_quotas	907	77	0
dc_users	10	0	0
outstanding_alerts	576	288	0

Cache：字典缓存类名

GES Requests：

GES Conflicts：

GES Releases：

Library Cache Activity (RAC)

Namespace	GES Lock Requests	GES Pin Requests	GES Pin Releases	GES Inval Requests	GES Invali- dations
ACCOUNT_STATUS	242	0	0	0	0
BODY	0	1,530,013	1,530,013	0	0
CLUSTER	74	74	74	0	0
DBLINK	246	0	0	0	0
EDITION	311	311	311	0	0
HINTSET OBJECT	186	186	186	0	0
INDEX	152,360	152,360	152,360	0	0
QUEUE	223	9,717	9,717	0	0
SCHEMA	255	0	0	0	0
SUBSCRIPTION	0	26	26	0	0
TABLE/PROCEDURE	275,215	3,023,083	3,023,083	0	0
TRIGGER	0	384,493	384,493	0	0

Namespace：library cache 的命名空间

GES Lock Requests：

GES Pin Requests：

GES Inval Requests：

GES Invali-dations：

Interconnect Ping Latency Stats

Ping latency of the roundtrip of a message from this instance to
target instances.
The target instance is identified by an instance number.
Average and standard deviation of ping latency is given in miliseconds
for message sizes of 500 bytes and 8K.
Note that latency of a message from the instance to itself is used as
control, since message latency can include wait for CPU

Target Instance	500B Ping Count	Avg Latency 500B msg	Stddev 500B msg	8K Ping Count	Avg Latency 8K msg	Stddev 8K msg
1	1,138	0.20	0.03	1,138	0.20	0.03
2	1,138	0.17	0.04	1,138	0.20	0.05
3	1,138	0.19	0.22	1,138	0.23	0.22
4	1,138	0.18	0.04	1,138	0.21	0.04

Target Instance：目标实例

500B Ping Count：

Avg Latency 500B msg：

Stddev 500B msg：

8K Ping Count：

Avg Latency 8K msg：

Stddev 8K msg：

Interconnect Throughput by Client

Throughput of interconnect usage by major consumers
All throughput numbers are megabytes per second

Used By	Send Mbytes/sec	Receive Mbytes/sec
Global Cache	0.10	0.20
Parallel Query	0.02	0.06
DB Locks	0.09	0.09
DB Streams	0.00	0.00
Other	0.02	0.01

Used By：主要消费者

Send Mbytes/sec：发送Mb/每秒

Receive Mbytes/sec：接收Mb/每秒

Interconnect Device Statistics

Throughput and errors of interconnect devices (at OS level)
All throughput numbers are megabytes per second

Device Name

IP Address

Public

Source

Send Mbytes/sec

Send Errors

Send Dropped

Send Buffer Overrun

Send Carrier Lost

Receive Mbytes/sec

Receive Errors

Receive Dropped

Receive Buffer Overrun

Receive Frame Errors

bondib0

192.168.10.8

cluster_interconnects parameter

0.00

Device Name：设备名称

IP Address：IP地址

Public：是否为公用网络

Source：来源

Send Mbytes/sec：发送MB/每秒

Send Errors：发送错误

Send Dropped：

Send Buffer Overrun：

Send Carrier Lost：

Receive Mbytes/sec：

Receive Errors：

Receive Dropped：

Receive Buffer Overrun：

Receive Frame Errors：

Dynamic Remastering Stats

times are in seconds
Affinity objects – objects mastered due to affinity at begin/end snap

Name	Total	per Remaster Op	Begin Snap	End Snap
remaster ops	29	1.00
remastered objects	40	1.38
replayed locks received	1,990	68.62
replayed locks sent	877	30.24
resources cleaned	0	0.00
remaster time (s)	5.0	0.17
quiesce time (s)	1.7	0.06
freeze time (s)	0.6	0.02
cleanup time (s)	0.7	0.02
replay time (s)	0.2	0.01
fixwrite time (s)	1.3	0.04
sync time (s)	0.5	0.02
affinity objects			365	367

Name：

Total：

Per Remaster Op：

Begin Snap：

End Snap：

Oracle RAC AWR指标含义

RAC相关指标

Dictionary Cache Stats (RAC)

Comments

Leave a Reply Cancel reply