AIX7环境上一次艰难的Oracle打补丁经历

系统环境

AIX ：7200-05-03-2148 Oracle：11.2.0.4 PSU: 11.2.0.4.201020（31718723） perl:5.28

问题一：AUTO patch

#/u01/app/11.2.0/grid/OPatch/opatch auto /tmp/31718723 错误信息如下：匹配mos 2516761.1 进行下一步relink操作

出师未捷，做relink，失败。

问题二：手工打补丁失败

执行perl rootcrs.pl -unlock出现报错。

复制代码

/usr/opt/perl5/lib/5.28.1/aix-thread-multi/POSIX.pm line 185.
 Unimplemented: POSIX::tmpnam() at crsconfig_lib.pm line 712.
 BEGIN failed--compilation aborted at crsconfig_lib.pm line 712.
 Compilation failed in require at ./rootcrs.pl line 305.

根据MOS文档Rootcrs.pl -Unlock Fails With Error Unimplemented: POSIX::tmpnam(): Use File::Temp Instead At /usr/perl5/5.26/lib/sun (文档 ID 2697590.1) 中说明在perl 5.22版本之后tmpnam() interface以及被删除，导致打补丁报错。

MOS解决办法 根据mos的建议，需要安装补丁27212837，不过由于也需要调用perl，也会出现/usr/opt/perl5/lib/5.28.1/aix-thread-multi/POSIX.pm line 185.的报错，出现了一个死循环，那么需要首先解决perl的问题。

复制代码

Apply patch 27212837
Download Link: https://updates.oracle.com/download/27212837.html
Request for patch if not available for particular platform/release.

临时解决： 用File::Temp替换qw(tmpnam)，包括crsgpnp.pm、crsinstall.pm、crsupgrade.pm、s_crsutils.pm，不过试了后发现crs无法启动。

复制代码

use File::Temp in the scripts instead of POSIX::tmpnam().
Replace use POSIX qw(tmpnam); with use File::Temp; in all the files, Then re-try
These are the files which have pragma use POSIX qw(tmpnam);
crsgpnp.pm
crsinstall.pm
crsupgrade.pm
s_crsutils.pm

问题三：降级perl版本无效

IBM厂商要来5.10版本的perl尝试手动应用补丁，发现补丁程序并没有使用我们指定的perl版本。然后我们也尝试用oracle自带的perl，写入环境变量，发现报错一样，也就是环境变量也没生效。 ++注：从小机厂商处了解到perl脚本因依赖比较多，无法直接卸载，也无降级一说。但是不同版本的perl脚本可以共存。没有影响。我们的例子中原本高版本的perl脚本没动。重新装了一个低版本5.10的perl。++

另辟蹊径：

既然环境变量不管用，那就使用oracle自带的perl并且强制写死路径，绕过操作系统的perl。

复制代码

/u01/app/11.2/grid/perl/bin/perl /u01/app/11.2.0//grid/crs/install/rootcrs.pl -unlock

rootcrs.pl -unlock这一步通过。

问题四：rootcrs.pl -patch报错

在最后一步rootcrs.pl -patch的时候报错install/crsconfig_lib.pm line 11821。

查看install/crsconfig_lib.pm 的11821代码，提示集群启动失败。

ASM没有启动成功，进程状态为Abnormal Termination,CLEANING

ASM的alert日志报错SMR is corrupted. Shut down and restart the instance to recreate it。

显然这是触发了BUG：AIX:11.2.0.4 :ASM&DB Instance Not Starting Up Due to SMR Corrupt After OCT 2020 PSU (文档 ID 2732507.1)

复制代码

APPLIES TO:
Oracle Database - Enterprise Edition - Version 11.2.0.4 to 11.2.0.4 [Release 11.2]
IBM AIX on POWER Systems (32-bit)
IBM AIX on POWER Systems (64-bit)
SYMPTOMS
The following errors reported in the ASM&DB alert logs while starting the ASM/DB instance after applying the OCT 2020 PSU.

ASM alert log:

Starting ORACLE instance (normal)
LICENSE_MAX_SESSION = 0
LICENSE_SESSIONS_WARNING = 0
Initial number of CPU is 32

Autotune of undo retention is turned on.

IMODE=BR
ILAT =0
LICENSE_MAX_USERS = 0
SYS auditing is disabled
NOTE: Volume support enabled
Starting up:
Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
With the Automatic Storage Management option.
ORACLE_HOME = <grid home>
System name: AIX
Node name: <node>
Release: 1
Version: 7
Machine: <Machine ID>
Using parameter settings in server-side spfile +DATA01/asm/asmparameterfile/registry.253.801309187
System parameters with non-default values:
large_pool_size = 12M
instance_type = "asm"
remote_login_passwordfile= "EXCLUSIVE"
asm_diskgroups = "FRA"
asm_diskgroups = "DATA01"
asm_power_limit = 1
diagnostic_dest = "<oracle base>"
Fri Nov 06 00:59:39 2020
WARNING: process PMON (ospid: 27132378) was unable to attach SMR.
SMR is corrupted. Shut down and restart the instance to recreate it.
PMON started with pid=2, OS id=27132378
Fri Nov 06 00:59:39 2020
WARNING: process PSP0 (ospid: 20512830) was unable to attach SMR.
SMR is corrupted. Shut down and restart the instance to recreate it.

Instance terminated by USER, pid = 20512820
Fri Nov 06 00:59:39 2020
Instance shutdown complete

在 AIX7 中应用 2020 年 10 月 DBPSU 后，SMR 文件不断损坏导致实例不能正常启动。

复制代码

BUG 32123313 - SMR FILE KEEPS GETTING CORRUPTED AFTER APPLYING OCT 2020 DBPSU IN AIX7
Bug 32109594 - ASM IS UNABLE TO START WITH SMR CORRUPTED ERROR POST OCT 2020 PSU APPLY

官方建议Download and apply the patch for bug 32109594 to GI/RDBMS homes.

复制代码

SOLUTION
Download and apply the patch for bug 32109594 to GI/RDBMS homes. In case the patch doesn't exist for your version, contact Oracle Support for the patch.

但是32109594是扩展服务，需要收费，很多客户没有权限下载。

解决办法：

其实有个变通的方法，既然官方说安装OCT 2020 DBPSU IN AIX7有问题，那我们我们尝试下安装上一个psu版本 JUL2020 (11.2.0.4.200714)。

安装成功。

作者介绍：云和恩墨资深Oracle dba，专注于数据库运维、架构和行业发展，有12年左右的金融、保险、政府、地税、运营商等业务关键型系统的运维经验，曾担任公司异常恢复东区接口人，负责紧急异常恢复工作，技术二线专家。目前负责PG、openGauss/MogDB运维、国产化MogDB数据库的推广工作。

本文由mdnice多平台发布