oracle中master实例,Oracle10g/11g RAC数据库中的Master实例、Owner实例和Past Image的概念PART3...

实验第33步：

下面来验证：

节选实例1上的/u01/app/oracle/admin/RDBA/udump/rdba1_ora_18378.trc：

/u01/app/oracle/admin/RDBA/udump/rdba1_ora_18378.trc

Oracle Database 10g Enterprise Edition  Release 10.2.0.1.0 - Production

With the Partitioning, Real Application  Clusters, OLAP and Data Mining options

ORACLE_HOME =  /u01/app/oracle/product/10.2.0/db_1

System name:Linux

Node name:node1.example.com

Release:2.6.18-164.el5xen

Version:#1 SMP Tue Aug 18 16:06:30 EDT 2009

Machine: i686

Instance name: RDBA1

Redo thread mounted by this instance: 1

Oracle process number: 25

Unix process pid: 18378, p_w_picpath:  oracle@node1.example.com (TNS V1-V3)

*** 2014-01-13 16:22:07.764

*** SERVICE NAME:(SYS$USERS) 2014-01-13  16:22:07.763

*** SESSION  ID:(125.1496) 2014-01-13 16:22:07.763

*****************************************************************

GLOBAL ENQUEUE  SERVICE DEBUG INFORMATION

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

Resource hash bucketcount

04

11

24

311

46

56

61

73

84

……

20392

20402

20416

20425

20432

20449

20452

20462

20473

Total resource count in hash buckets: 8213

***************** End of lkdebug output  *************************

*** 2014-01-13 16:26:26.465

*****************************************************************

GLOBAL ENQUEUE SERVICE DEBUG INFORMATION

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

node# 0, #nodes 2, state 4, msgver 4,  rcvver 0 validver 4

valid_domain 1

sync acks 0x000000000000000000000000000000000

Resource freelist #0 len 28410 lwm 2893  add 241108 rem 212698

Resource freelist #1 len 28471 lwm 3306  add 241942 rem 213471

LMS0:

Hash buckets log2(11)

Bucket# 0 #res 0

Bucket# 1 #res 0

Bucket# 2 #res 0

Bucket# 3 #res 0

Bucket# 4 #res 0

Bucket# 5 #res 0

Bucket# 6 #res 0

Bucket# 7 #res 0

……

atch buckets log2(6)

GCS shadow freelist #0 len 29067 lwm 7451  add 88332 rem 59265

GCS shadow freelist #1 len 29097 lwm 7257  add 88862 rem 59765

files in affinity vector:

* >> PT table contents ---:

pt table bucket = 1

pkey 4294950913, stat 0, masters[32767,  0->0], reminc 2, RM# 1 flg 0x0

pt table bucket = 2

pkey 4294950914, stat 0, masters[32767,  0->0], reminc 2, RM# 1 flg 0x0

pt table bucket = 3

pkey 4294950915, stat 0, masters[32767,  0->0], reminc 2, RM# 1 flg 0x0

pkey 52533, stat 0, masters[0, 1->1],  reminc 4, RM# 6 flg 0x0←手工 remaster之前， oradebug lkdebug –k的输出，代表master实例是2([0,  1->1]中的1->1表示实例2)，上一任 master是实例1([0, 1->1]中的0表示上一任master实例1)

* kjilpkey = 0

***************** End of lkdebug output  *************************

*** 2014-01-13 16:27:14.981

*****************************************************************

GLOBAL ENQUEUE SERVICE DEBUG INFORMATION

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

***************** End of lkdebug output  *************************

Latch buckets log2(6)

GCS shadow freelist #0 len 7487 lwm 7451  add 88336 rem 80849

GCS shadow freelist #1 len 7569 lwm 7257  add 88863 rem 81294

files in affinity vector:

* >> PT table contents ---:

pkey 4294950932, stat 0, masters[32767,  1->1], reminc 4, RM# 4 flg 0x0

pt table bucket = 3381

pkey 52533, stat 0, masters[1, 0->0],  reminc 4, RM# 7 flg 0x0←手工 remaster之后， oradebug lkdebug –k的输出，代表master实例是1([1,  0->0]中的0->0表示实例1)，上一任 master是实例1([1, 0->0]中的1表示上一任master实例2)

* kjilpkey = 1

***************** End of lkdebug output  *************************

trace文件已经说明：实例1重新夺回了对象52533所有的块的mastership。

select*from  myviewwhere"MASTER_Instance"=2 ;

无输出。这就对了，因为都被实例1master了。

select* from v$gcspfmaster_info where object_id=52533;

FILE_IDOBJECT_IDCURRENT_MASTERPREVIOUS_MASTERREMASTER_CNT

10525330↖代表master实例是11↖代表上一任master是实例20

select drms from X$KJDRMAFNSTATS;

DRMS

15

DRM为4+1=5，验证了此刻又发生了一次Remaster(第3次Remaster)。

再验证：

进入实例1的/u01/app/oracle/admin/RDBA/bdump:

执行：grep-r"pkey  52533"./

输出：

./rdba1_lmd0_7002.trc:Transfer pkey 52533 to node 0

./rdba1_lmd0_7002.trc:Begin DRM(5) - transfer pkey  52533 to 0 oscan 0.0

./rdba1_lmd0_7002.trc:Rcvd DRM(6) Transfer pkey  52533 from 0 to 1 oscan 0.0

./rdba1_lmd0_7002.trc:Transfer pkey 52533 to node 0

./rdba1_lmd0_7002.trc:Begin DRM(7) -  transfer pkey 52533 to 0 oscan 0.0

进入实例2的/u01/app/oracle/admin/RDBA/bdump:

执行：grep-r"pkey  52533"./

输出：

./rdba2_lms0_7034.trc: GCS CLIENT 0x233f71b0,2  sq[(nil),(nil)] resp[(nil),0x185.40000] pkey 52533

./rdba2_lms0_7034.trc:pkey 52533

./rdba2_lms0_7034.trc: GCS CLIENT 0x233f71b0,2  sq[(nil),(nil)] resp[(nil),0x185.40000] pkey 52533

./rdba2_lms0_7034.trc:pkey 52533

./rdba2_lmd0_7032.trc:Rcvd DRM(5) Transfer pkey  52533 to 0 oscan 1.1

./rdba2_lmd0_7032.trc:Transfer pkey 52533 to node 1

./rdba2_lmd0_7032.trc:Begin DRM(6) - transfer pkey  52533 to 1 oscan 0.0

./rdba2_lmd0_7032.trc:Rcvd  DRM(7) Transfer pkey 52533 from 1 to 0 oscan 0.0