ITPUB论坛-中国最专业的IT技术社区

 
 注册
热搜:
查看: 251|回复: 1

[笔记] 12C强制ASM DISK 大小相同以及exadata上该问题 处理方式

[复制链接]
论坛徽章:
9
娜美
日期:2017-06-26 15:18:15目光如炬
日期:2017-08-20 22:00:00目光如炬
日期:2017-08-27 22:00:01火眼金睛
日期:2017-09-01 17:00:07目光如炬
日期:2017-09-03 22:00:01火眼金睛
日期:2017-09-30 22:00:01目光如炬
日期:2017-11-26 22:00:00目光如炬
日期:2017-12-17 22:00:00目光如炬
日期:2018-01-14 22:00:00
跳转到指定楼层
1#
发表于 2017-12-7 16:49 | 只看该作者 回帖奖励 |倒序浏览 |阅读模式


目前就职海天起点,服务于电力行业,致力于帮助客户解决生产过程中出现的问题,提高生产效率, 爱好书法,周易!愿结交志同道合之士!共同进步! 微信号:sunyunyi_sun

我们经常发现在运的ASM 磁盘容量大小不同,因为存储工程师和DBA没有很好的沟通,主要原因在DBA没有严格把关,有些系统磁盘容量100g,500g,200g各种大小都有,导致磁盘空间
不能最大利用,资源浪费,10G环境中还存在一些BUG,从12c版本开始ORACLE强制ASM 磁盘大小相同,即使不是12C强烈建议ASM磁盘大小相同。下面我们看看ORACLE的限制。

1) Starting on 12.1.0.2 ASM release, this ASM constraint/validation is available:

15410, 00000, "Disks in disk group %s do not have equal size."
// *Cause: The disks in the diskgroup were not of equal size.
// *Action: Ensure that all disks in the diskgroup are of equal size. If
//          adding new disks to the diskgroup, their size must be equal to
//          the size of the existing disks in the diskgroup. If resizing, all
//          disks in the diskgroup must be resized to the same size.


2) Disks with uneven capacity can create allocation problems (e.g. "ORA-15041: diskgroup space exhausted" that prevent full use of all of the available storage in the

failgroup /diskgroup.  

3) This validation/constraint ensure that all disks in the same diskgroup have the same size, doing so provides more predictable overall performance and space

utilization.

4) If the disks are the same size, then ASM spreads the files evenly across all of the disks in the diskgroup. This allocation pattern maintains every disk at the same

capacity level and ensures that all of the disks in a diskgroup have the same I/O load. Because ASM load balances workload among all of the disks in a diskgroup,

different ASM disks should not share the same physical drive.

5) This ASM new feature is enabled by default on '12.1.0.2' Grid Infrastructure/ASM release and onwards.

从 12.1.0.2 开始 oracle 强制 ASM disk 必须大小相同。下面给出了大小不相同时的处理方案:

1) Associated diskgroup was created using the 'COMPATIBLE.ASM'='12.1.0.2.0' attribute and the disks size was =10,239M:

[grid@asmlnx1 ~]$ sqlplus "/as sysasm"

SQL*Plus: Release 12.1.0.2.0 Production on Sat Oct 25 16:27:50 2014

Copyright (c) 1982, 2014, Oracle.  All rights reserved.


Connected to:
Oracle Database 12c Enterprise Edition Release 12.1.0.2.0 - 64bit Production
With the Automatic Storage Management option

SQL> CREATE DISKGROUP DATA12C NORMAL REDUNDANCY  
FAILGROUP FG2 DISK 'ORCLISK3_SAN_B' SIZE 10239M  
FAILGROUP FG1 DISK 'ORCLISK3_SAN_A' SIZE 10239M  
ATTRIBUTE 'compatible.asm'='12.1.0.2.0','compatible.rdbms'='12.1.0.2.0','au_size'='1M';

Diskgroup created.


2) But the new candidate disks have a different size (e.g. 20,489M) reported at OS level as follows:

GROUP_NUMBER         DISK_NUMBER         MOUNT_S         HEADER_STATU         MODE_ST         STATE         REDUNDA         OS_MB         PATH
0        4        CLOSED         PROVISIONED        ONLINE         NORMAL         UNKNOWN         20489        ORCL:SAN_A_DISK_2
0        17        CLOSED         PROVISIONED        ONLINE         NORMAL         UNKNOWN         20489        ORCL:SAN_B_DISK_3



3) You have 2 options to solve this problem:

Option #1:

Add new disks to the associated diskgroup with the same size (e.g. 10,239M implicitly reported at OS level) as the existing disk member disks as follows:

ROUP_NUMBER         DISK_NUMBER         MOUNT_S         HEADER_STATU         MODE_ST         STATE         REDUNDA         OS_MB         PATH
0        4        CLOSED         PROVISIONED        ONLINE         NORMAL         UNKNOWN         10239        ORCL:SAN_A_DISK_6
0        17        CLOSED         PROVISIONED        ONLINE         NORMAL         UNKNOWN         10239        ORCL:SAN_B_DISK_5



[grid@asmlnx1 ~]$ sqlplus "/as sysasm"

SQL*Plus: Release 12.1.0.2.0 Production on Sat Oct 25 17:07:53 2014

Copyright (c) 1982, 2014, Oracle.  All rights reserved.


Connected to:
Oracle Database 12c Enterprise Edition Release 12.1.0.2.0 - 64bit Production
With the Automatic Storage Management option

SQL> alter diskgroup DATA12c add failgroup F1 disk 'ORCL:SAN_A_DISK_6' failgroup F2 disk 'ORCL:SAN_B_DISK_5';

Diskgroup altered.


Option #2:

Use the original bigger disks (e.g. 20,489M), but explicitly restrict/define the smaller/even required disk size (e.g. 10,239M) at ASM level as follows:





GROUP_NUMBER         DISK_NUMBER         MOUNT_S         HEADER_STATU         MODE_ST         STATE         REDUNDA         OS_MB         PATH
0        4        CLOSED         PROVISIONED        ONLINE         NORMAL         UNKNOWN         20489        ORCL:SAN_A_DISK_2
0        17        CLOSED         PROVISIONED        ONLINE         NORMAL         UNKNOWN         20489       
ORCL:SAN_B_DISK_3



[grid@asmlnx1 ~]$ sqlplus "/as sysasm"

SQL*Plus: Release 12.1.0.2.0 Production on Sat Oct 25 17:07:53 2014

Copyright (c) 1982, 2014, Oracle.  All rights reserved.


Connected to:
Oracle Database 12c Enterprise Edition Release 12.1.0.2.0 - 64bit Production
With the Automatic Storage Management option

SQL> alter diskgroup DATA12c add failgroup F1 disk 'ORCL:SAN_A_DISK_2' size 10239M failgroup F2 disk 'ORCL:SAN_B_DISK_3' size 10239M;

Diskgroup altered.

很显然这种方式浪费资源


exadata 同样存在该问题:
Oracle Database - Enterprise Edition - Version 12.1.0.2 to 12.1.0.2 [Release 12.1]
Oracle Exadata Storage Server Software - Version 11.1.0.3.0 to 12.1.2.2.3 [Release 11.1 to 12.1]

SQL> alter diskgroup RECO_ASMCLOUD add disk 'o/*/RECO_CD*ASMCLOUDCEL01' rebalance power 32;
alter diskgroup RECO_ASMCLOUD add disk 'o/*/RECO_CD*ASMCLOUDCEL01' rebalance power 32
*
ERROR at line 1:
ORA-15032: not all alterations performed
ORA-15410: Disks in disk group RECO_ASMCLOUD do not have equal size.

处理方法

1:在大小不一致cell节点修改磁盘大小(下面以 asmcloudcel01 cell )

# cellcli -e alter griddisk
RECO_CD_00_ASMCLOUDCEL1,
RECO_CD_01_ASMCLOUDCEL1,
RECO_CD_02_ASMCLOUDCEL1,
RECO_CD_03_ASMCLOUDCEL1,
RECO_CD_04_ASMCLOUDCEL1,
RECO_CD_05_ASMCLOUDCEL1,
RECO_CD_06_ASMCLOUDCEL1,
RECO_CD_07_ASMCLOUDCEL1,
RECO_CD_08_ASMCLOUDCEL1,
RECO_CD_09_ASMCLOUDCEL1,
RECO_CD_10_ASMCLOUDCEL1,
RECO_CD_11_ASMCLOUDCEL1
size=<New value>G;

2: Add the new griddisks back from the asmcloudcel01 cell to the affected diskgroup (e.g. RECO_ASMCLOUD):

SQL> alter diskgroup RECO_ASMCLOUD add failgroup asmcloudcel01 disk
'o/192.999.99.3/RECO_CD_09_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_10_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_05_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_07_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_01_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_06_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_11_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_00_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_02_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_03_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_04_asmcloudcel01' SIZE 197632 M,
'o/192.999.99.3/RECO_CD_08_asmcloudcel01' SIZE 197632 M rebalance power 32;

3:所有 cell 执行:
# cellcli -e alter griddisk
RECO_CD_00_ASMCLOUDCEL1,
RECO_CD_01_ASMCLOUDCEL1,
RECO_CD_02_ASMCLOUDCEL1,
RECO_CD_03_ASMCLOUDCEL1,
RECO_CD_04_ASMCLOUDCEL1,
RECO_CD_05_ASMCLOUDCEL1,
RECO_CD_06_ASMCLOUDCEL1,
RECO_CD_07_ASMCLOUDCEL1,
RECO_CD_08_ASMCLOUDCEL1,
RECO_CD_09_ASMCLOUDCEL1,
RECO_CD_10_ASMCLOUDCEL1,
RECO_CD_11_ASMCLOUDCEL1
size=<New value>G;

4:Finally, please resize up the ASM disks in the affected diskgroup
SQL> alter diskgroup RECO_ASMCLOUD resize all rebalance power 32;


从上面的信息可以看到ASM DISK 大小相同可以带来的好处:数据均衡,均衡 I/O负载,最大的可用性。




论坛徽章:
69
山治
日期:2017-09-11 19:15:33处女座
日期:2015-11-27 12:27:01秀才
日期:2015-11-23 10:17:19秀才
日期:2015-11-23 09:57:36金牛座
日期:2016-04-01 18:05:22秀才
日期:2015-11-11 10:22:49摩羯座
日期:2015-08-20 16:22:52秀才
日期:2015-08-31 13:02:54秀才
日期:2015-12-25 15:31:10巨蟹座
日期:2015-11-04 12:43:04
2#
发表于 2017-12-20 12:11 | 只看该作者
这个特性完全是为Exadata量身定制的。非Exadata环境,看到的LUN过了那么多层,很难保证不同的LUN来自不同的RAID组,这样做其实也没啥意义。尤其EMC的存储。

使用道具 举报

回复

您需要登录后才可以回帖 登录 | 注册

本版积分规则

DTCC2018购票6.8折优惠进行时

中国数据库技术大会是国内数据库及大数据领域规模最大、最受欢迎的技术交流盛会。 2018年5月10-12日,第九届中国数据库技术大会将如约而至。本届大会以“数领先机•智赢未来”为主题,设定2大主会场及20个技术专场,邀请来自国内外互联网、金融、教育等行业百余位技术专家,共同探讨Oracle、MySQL、NoSQL、大数据等领域的前瞻性热点话题与技术。
----------------------------------------
优惠时间:2018年2月13日前

TOP技术积分榜 社区积分榜 徽章 团队 统计 知识索引树 积分竞拍 文本模式 帮助
  ITPUB首页 | ITPUB论坛 | 数据库技术 | 企业信息化 | 开发技术 | 微软技术 | 软件工程与项目管理 | IBM技术园地 | 行业纵向讨论 | IT招聘 | IT文档 |
  | | |
CopyRight 1999-2011 itpub.net All Right Reserved. 北京盛拓优讯信息技术有限公司版权所有 联系我们 网站律师 隐私政策 知识产权声明
 北京市公安局海淀分局网监中心备案编号:11010802021510 广播电视节目制作经营许可证:编号(京)字第1149号
  
快速回复 返回顶部 返回列表
北京赛车pk10 申博官网 北京赛车pk10 六台宝典现场开奖 北京赛车pk10历史记录 北京赛车开奖 北京赛车预测 949494开奖结果今晚 威尼斯人线上娱乐 手机投注平台 幸运28投注技巧 pk10助赢软件 北京pk10百度鼎盛彩票网 北京赛车聚彩 北京pk10如何稳杀3码 pk10稳赢方法 pk10定位计划 pk10包赢计划群