Typical ORA-12547 error handling

ORA-12547 error handling today, the leader told me, let me help a colleague to install Oracle, he encountered a problem in the installation process, error, unable to continue to install. I

ORA-12547 error handling today, the leader told me, let me help a colleague to install Oracle, he encountered a problem in the installation process, error, unable to continue to install. I

ORA-12547 error handling

Today, the supervisor told me that I asked a colleague to install Oracle. he encountered a problem during the installation and reported an error, so he could not continue the installation.

1. ora-12547 error due to NLS_LANG setting error
I looked at it and found that the ORA-12547 tns lost connect error immediately occurred when DBCA was used in the AIX5.3 system to build the database, then, the system environment is checked in detail. asynchronous IO is enabled, and the HOST file is also normal. The parameter file does not seem to be too large, the ulimit limit is changed to-1, and no obvious problems are found. The installation log shows that the operating system environment is not satisfied. At that time, I couldn't find out why I deleted the previously installed DBSOFT and re-installed the DBSOFT database. No error was reported during DBSOFT installation, we had to check the problem again and found that [db: oracle] $ cat ~ in the oracle user environment configuration file ~ /. The NLS_LANG = ZHS16GBK parameter is found in the profile. Later I think this parameter is not required, and it may cause an error during installation. Then I disable it. The parameter file is as follows: $ cat. profile

PATH =/usr/bin:/etc:/usr/sbin:/usr/ucb: $ HOME/bin:/usr/bin/X11:/sbin :.
Export PATH
If [-s "$ MAIL"] # This is at Shell startup. In normal
Then echo "$ MAILMSG" # operation, the Shell checks
Fi # periodically.
ORACLE_BASE =/oracle
ORACLE_SID = orcl
ORACLE_HOME = $ ORACLE_BASE/app/10.2.0/db_1
# NLS_LANG = ZHS16GBK
PATH = $ ORACLE_HOME/bin: $ PATH:
Export ORACLE_BASE ORACLE_SID ORACLE_HOME PATH
$
After disabling the database, the database is created again in DBCA, and the fault does not recur. The database is created successfully.

2. This problem occurs in dbca database creation. After searching on the Internet, it is basically caused by the following four reasons:
1. asynchronous IO causes dbca to report ORA-12547
Solution:
AIX:
A. smitty chaio (only supported by the AIX 5 Series)
B. Set STATE to be configured at system restart to available.
C. Run DBCA again and create it successfully.
HP-UNIX For reference (Metalink Note: 139272.1)

A. Create an aio character device
#/Sbin/mknod/dev/async c 101 0x0
# Chown oracle: dba/dev/async
# Chmod 660/dev/async

B. Use sam to adjust Kernel Parameters
-> Kernel Configuration
-> Drivers
-> Select 'asyncdsk'
-> Actions> Add Driver to Kernel. (The pending state changes from out to in)
-> Kernel Configuration
-> Retriable Parameters
-> Max_async_ports
Set to the 'processs' value in the initialization file + number of oracle background processes (according to v $ bgprocess, the maximum number of background processes for a single instance is 78)

C. Use sam to set aio_max_ops
-> Kernel Configuration
-> Retriable Parameters
-> Aio_max_ops
-> The default value is 2048.

In the preceding configuration, the core of the host needs to be reconstructed and restarted.
D. The oracle user must have the MLOCK permission.
A. Set the MLOCK permission for the dba group.
Roolt # setprivgrp dba MLOCK

B. To avoid failure after restart, you need to set:
Add the following lines to vi/etc/privgroup:
Dba MLOCK
E. Check whether aio has taken effect.
# Fuser/dev/async
If a process is listed, it indicates that aio has taken effect.



2. The default value of ulimit is not modified.
Solution:
Linux
A. Modify the limits parameter file and add the following content:
Vi/etc/security/limits. conf
Oracle soft nproc 2047
Hard nproc 16384
Oracle soft nofile 1024
Oracle hard nofile 65536

AIX
A. Modify the limits parameter file and add the following content at the end: (Note:-1 indicates unlimited use)
Vi/etc/security/limits
Default:
Fsize =-1
Core = 2097151
Cpu =-1
Data = 262144
Rss = 65536
Stacks = 65536
No files = 2000
Root:
Daemon:
Bin:
Sys:
Adm:
Uucp:
Guest:
Nobody:
Lpd:
Oracle:
Data =-1
Stack =-1

3. The SELinux configuration in Linux is incorrect. Solution
A. Use the root user to log on to the system and execute setenforce 0 (no need to restart)
B. Modify selinux = "" in the/etc/SELINUX/config file to disabled and restart

4. libaio (x86_64) is not installed(Like AIX, asynchronous IO is enabled and this package is used in Linux)
Solution:
A. Insert the operating system installation disc
B. log on to the system as the root user and mount the operating system installation disc:
# Mount-t iso9660/dev/cdrom/mnt
# Cd/mnt/Server
C. Check and install the libaio package and run the following command:
# Rpm-qa | grep-I ^ libaio
# Rpm-ivh libaio-*. i386.rpm
# Rpm-ivh libaio-*. x86_64.rpm
D. Check whether the package has been installed successfully.
# Rpm-qa | grep ^ libaio
E. Run the installation again to solve the problem.

For more information about Oracle, see the Oracle topic page? Tid = 12