[20210902]为什么会使用多个共享内存段.txt

--//昨天测试kernel.sem = SEMMSL SEMMNS SEMOPM SEMMNI参数时遇到的问题,做一个记录.

1.环境:
--//为了反复测试,我建立一个实例,没有数据文件.
$ cat /tmp/test.ora
test.__db_cache_size=1258291200
test.__oracle_base='/u01/app/oracle'#ORACLE_BASE set from environment
test.__shared_pool_size=1048576000
*.db_name='test'
*.processes=200
*.sga_target=2500M
*.undo_management='auto'

--//设置环境变量:
$ export ORACLE_SID=test

# egrep "^kernel.s[he]|^vm.nr_" /etc/sysctl.conf
kernel.shmmax = 68719476736
kernel.shmall = 4294967296
kernel.shmmni = 4096
kernel.sem = 102 26112 204 128
vm.nr_hugepages = 404
vm.nr_overcommit_hugepages = 1300

# sysctl -p

--//kernel.shmmax=68719476736, 68719476736/1024/1024/1024 = 64G,理论sga_target=2500M,11g下正常建立3个共享内存段.
--//可是实际情况确实6个.通过例子说明.

2.测试:
SYS@test> startup force nomount pfile='/tmp/@.ora';
ORACLE instance started.
Total System Global Area 2622255104 bytes
Fixed Size                  2256112 bytes
Variable Size            1124074256 bytes
Database Buffers         1476395008 bytes
Redo Buffers               19529728 bytes

$ ipcs -m

------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status
0x00000000 45645826   oracle    640        33554432   18
0x00000000 45678595   oracle    640        654311424  18
0x00000000 45711364   oracle    640        134217728  18
0x00000000 45744133   oracle    640        16777216   18
0x00000000 45776902   oracle    640        1795162112 18
0x8a931fb8 45809671   oracle    640        2097152    18
--//使用6个共享内存段.

$ ipcs -m -u
------ Shared Memory Status --------
segments allocated 8
pages allocated 643681
pages resident  79750
pages swapped   0
Swap performance: 0 attempts     0 successes
--//segments allocated 8 ,实际上有2个被root用户使用.如果仔细看alert日志:

Starting ORACLE instance (normal)
************************ Large Pages Information *******************
Per process system memlock (soft) limit = 60 GB
Total Shared Global Region in Large Pages = 802 MB (31%)
Large Pages used by this instance: 401 (802 MB)
Large Pages unused system wide = 3 (6144 KB)
Large Pages configured system wide = 404 (808 MB)
Large Page size = 2048 KB

RECOMMENDATION:
  Total System Global Area size is 2514 MB. For optimal performance,
  prior to the next instance restart:
  1. Increase the number of unused large pages by
 at least 853 (page size 2048 KB, total size 1706 MB) system wide to
  get 100% of the System Global Area allocated with large pages
********************************************************************

--//噢,我给hugepages仅仅404,完全不能满足需求,但是我定义了vm.nr_overcommit_hugepages = 1300,1300*2M=2600M,完全满足不需要
--//建立使用这么多的共享内存段.为什么呢?
--//仔细看http://blog.itpub.net/267265/viewspace-2145481/=>0927hugepages与nr_overcommit_hugepages
--//才发现要定义use_large_pages=only才可以使用.

$ cat /tmp/test.ora
test.__db_cache_size=1258291200
test.__oracle_base='/u01/app/oracle'#ORACLE_BASE set from environment
test.__shared_pool_size=1048576000
*.db_name='test'
*.processes=200
*.sga_target=2500M
*.undo_management='auto'
*.use_large_pages=only

SYS@test> startup force nomount pfile='/tmp/@.ora';
ORACLE instance started.
Total System Global Area 2622255104 bytes
Fixed Size                  2256112 bytes
Variable Size            1124074256 bytes
Database Buffers         1476395008 bytes
Redo Buffers               19529728 bytes

$ ipcs -m
------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status
0x00000000 45875202   oracle    640        33554432   18
0x00000000 45907971   oracle    640        2600468480 18
0x8a931fb8 45940740   oracle    640        2097152    18

$ ipcs -m -u
------ Shared Memory Status --------
segments allocated 5
pages allocated 643681
pages resident  131681
pages swapped   0
Swap performance: 0 attempts     0 successes

--//这样就对了.有时候陷入困境,思维一下子跳不出来.
--//alert.log
************************ Large Pages Information *******************
Parameter use_large_pages = only
Per process system memlock (soft) limit = 60 GB

Total Shared Global Region in Large Pages = 2514 MB (100%)

Large Pages used by this instance: 1257 (2514 MB)
Large Pages unused system wide = 0 (0 KB)
Large Pages configured system wide = 1257 (2514 MB)
Large Page size = 2048 KB
********************************************************************

# grep -i pages /proc/meminfo
AnonPages:        277180 kB
AnonHugePages:     86016 kB
HugePages_Total:    1257
HugePages_Free:     1000
HugePages_Rsvd:     1000
HugePages_Surp:      853
Hugepagesize:       2048 kB

--//HugePages_Total: 1257,从vm.nr_overcommit_hugepages借用了853,853+404 = 1257.
--//HugePages_Total-HugePages_Free+HugePages_Rsvd 就是 目前实例需要的页面数量.