Oracle Database RAC 11.2.0.3 for AIX6.1TL7安装记录(3)

        下面是再一次安装Oracle 11.2.0.3 RAC Database for  AIX6.1 TL7遇到问题的记录，之前还有两篇记录文章：

《Oracle Database RAC 11.2.0.3 for AIX6.1TL7安装记录(1)》： http://space.itpub.net/23135684/viewspace-733990

《Oracle Database RAC 11.2.0.3 for AIX6.1TL7安装记录(2)》： http://space.itpub.net/23135684/viewspace-742017

《在AIX上运行RAC时网络方面的一些最佳经验》： https://blogs.oracle.com/Database4CN/entry/%E5%9C%A8aix%E4%B8%8A%E8%BF%90%E8%A1%8Crac%E6%97%B6%E7%BD%91%E7%BB%9C%E6%96%B9%E9%9D%A2%E7%9A%84%E4%B8%80%E4%BA%9B%E6%9C%80%E4%BD%B3%E7%BB%8F%E9%AA%8C


问题一：
        报错截图如下：

 
        在安装Grid Infrastructure的时候，图形界面会调用CVU验证安装环境，其中在验证网络参数的时候会出现验证失败的情况，但在操作系统层面通过no -a | grep ipqmaxlen命令验证相应参数已是正确配置。问题的解决办法如下：
PRVE-0273 : The value of network parameter "udp_sendspace" for interface "en0" is not configured to the expected value on node "racnode1" [ID 1373242.1]
修改时间: 2012-3-7 类型: REFERENCE 状态: MODERATED 优先级: 3

In this Document
  Purpose
  PRVE-0273 : The value of network parameter "udp_sendspace" for interface "en0" is not configured to the expected value on node "racnode1"
      
     bug 13077654 - AIX specific
     bug 13531373 - AIX specific 

---

This document is being delivered to you via Oracle Support's Rapid Visibility (RaV) process and therefore has not been subject to an independent technical review.

Applies to:

Oracle Server - Enterprise Edition - Version: 11.2.0.1 and later   [Release: 11.2 and later ]
Information in this document applies to any platform.

Purpose

This note lists typical causes and solutions for the following cluvfy error:

PRVE-0273 : The value of network parameter "rfc1323" for interface "en2" is not configured to the expected value on node "racnode1"

PRVE-0273 : The value of network parameter "udp_sendspace" for interface "en0" is not configured to the expected value on node "racnode1"

bug 13077654 - AIX specific

On AIX, runInstaller complains network parameter setting: ipqmaxlen, rfc1323, sb_max, tcp_sendspace, udp_sendspace, udp_recvspace

INFO: *********************************************
INFO: Network parameter -  rfc1323: Checks if the network parameter is set correctly on the system
INFO: Severity:IGNORABLE
INFO: OverallStatus:VERIFICATION_FAILED
INFO: -----------------------------------------------
INFO: Verification Result for Node:racnode1
INFO: Expected Value:1
INFO: Actual Value:en2=0
INFO: Error Message: PRVE-0273 : The value of network parameter "rfc1323" for interface "en2" is not configured to the expected value on node "racnode1".[Expected="1"; Found="en2=0"]

Manually verified with "ifconfig" and "/usr/sbin/no", the setting is as expected

This bug is fixed in 12.1 and onward

The workaround is to create a symbolic as root:

# ln -s /usr/sbin/no /etc/no

bug 13531373 - AIX specific

On AIX, runInstaller complains network parameter setting even when they are bigger than required: 

INFO: *********************************************
INFO: Network parameter -  tcp_sendspace: Checks if the network parameter is set correctly on the system
INFO: Severity:IGNORABLE
INFO: OverallStatus:VERIFICATION_FAILED
INFO: -----------------------------------------------
INFO: Verification Result for Node:racnode1
INFO: Expected Value:1
INFO: Actual Value:en2=0
INFO: Error Message: PRVE-0273 : The value of network parameter "tcp_sendspace" for interface "en10" is not configured to the expected value on node "racnode1".[ Expected=" 65536"; Found="en10= 262144"]

As you can see, the expected value is 65536, and the current value is 262144 is satisfies the requirement.

The fix is included in 11.2.0.3 GI PSU2, 11.2.0.4 and above, the error can be ignored.

问题二：
        报错截图如下：


        在安装Grid Infrastrcture的时候，根据提示在第一个节点执行root.sh脚本出现如上的错误信息（Failed to write the checkpoint:'' with status:FAIL.Error code is 256），问题的解决办法如下：
AIX: 11gR2 Grid Infrastructure Installation, root.sh Error: Failed to write the checkpoint:'' with status:FAIL.Error code is 256 [ID 1382505.1]
修改时间: 2011-12-5 类型: PROBLEM 状态: PUBLISHED 优先级: 3

In this Document
  Symptoms
  Cause
  Solution

---

Applies to:

Oracle Server - Enterprise Edition - Version: 11.2.0.1 and later   [Release: 11.2 and later ]
IBM AIX on POWER Systems (64-bit)

Symptoms

Trying to install Grid Infrastructure 11.2.0.3, root.sh fails with:

# root.sh
..
User ignored Prerequisites during installation
Failed to write the checkpoint:'' with status:FAIL.Error code is 256
Undefined subroutine &crsconfig_lib::dieformat called at /oracle/app/11.2.0.3/grid/crs/install/crsconfig_lib.pm line 6135.

rootcrs_.log shows:

2011-11-23 03:43:20: User ignored Prerequisites during installation
2011-11-23 03:43:24: ###### Begin DIE Stack Trace ######
2011-11-23 03:43:24: Package File Line Calling
2011-11-23 03:43:24: --------------- -------------------- ---- ----------
2011-11-23 03:43:24: 1: main rootcrs.pl 375 crsconfig_lib::dietrap
2011-11-23 03:43:24: 2: crsconfig_lib crsconfig_lib.pm 6135 main::__ANON__
2011-11-23 03:43:24: 3: crsconfig_lib crsconfig_lib.pm 6640 crsconfig_lib::set_file_perms
2011-11-23 03:43:24: 4: main rootcrs.pl 457 crsconfig_lib::run_env_setup_modules
2011-11-23 03:43:24: ####### End DIE Stack Trace #######
..
2011-11-23 03:43:24: Failed to write the checkpoint:'' with status:FAIL.Error code is 256

Cause

The problem is caused by clusterware library pointing to non-exist Vendor clusterware library, eg:

$ ls -l /oracle/app/11.2.0.3/grid/lib/libskgxn*
lrwxrwxrwx 1 grid oinstall 33 Nov 23 03:08 /oracle/app/11.2.0.3/grid/lib/libskgxn2.so -> /opt/ORCLcluster/lib/libskgxn2.so
-rwxr-xr-x 1 grid oinstall 159806 Oct 20 23:55 /oracle/app/11.2.0.3/grid/lib/libskgxnr.a
lrwxrwxrwx 1 grid oinstall 33 Nov 23 09:38 /oracle/app/11.2.0.3/grid/lib/libskgxnr.so -> /opt/ORCLcluster/lib/libskgxnr.so

$ ls -l /opt/ORCLcluster
ls: 0653-341 The file /opt/ORCLcluster does not exist.

This is caused by HACMP executable is not removed cleanly when HACMP is deinstalled. When HACMP is installed, it installs the directory /usr/sbin/cluster/utilities along with others. Oracle OUI depends on /usr/sbin/cluster/utilities/cldomain to determine if vendor clusterware exists. If yes, then a symlink of $GRID_HOME/lib/libskgxn2.so will be created pointing to /opt/ORCLcluster/lib/libskgxn2.so (so does libskgxnr.so). /opt/ORCLcluster directory is setup during rootpre.sh if vendor cluster is presented.

In this case, HACMP was first installed, then Veritas software was installed, it caused the /usr/sbin/cluster/utilities/cldomain became a symlink pointing to Veritas clusterware:

$ ls -l /usr/sbin/cluster/utilities/cldomain
lrwxrwxrwx    1 root     system           29 Sep 21 13:54 /usr/sbin/cluster/utilities/cldomain -> /opt/VRTSvcs/rac/bin/cldomain

When HACMP was deinstalled later, it removed all other files but left this symlink cldomain, causing Oracle considered vendor clusterware exists and created the symlink of libskgxn2.so and libskgxnr.so during link libraries phase in OUI installation. Further leads to root.sh failure.

Solution

1. When deinstalling vendor clusterware, make sure all associated files are removed. In this case, remove the symlink /usr/sbin/cluster/utilities/cldomain

2. Clean up the failed GI installation via $GRID_HOME/deinstall/deinstall command or clean up manually follow  DOCUMENT 1364419.1

3. Reinstall Grid Infrastructure 


--end--

1.jpg

2.jpg

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/23135684/viewspace-756012/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/23135684/viewspace-756012/