ORA-29701 from 10.2 DB after ASM upgraded to 11.2 in HP-UX Itanium_1323826.1

rongshiyuan發表於2014-10-14

ORA-29701 from 10.2 databases after ASM upgraded to 11.2 in HP-UX Itanium (文件 ID 1323826.1)


In this Document

Symptoms
Cause
Solution
References


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 Database - Enterprise Edition - Version 11.2.0.1 and later
Information in this document applies to any platform.
***Checked for relevance on 05-Sep-2013***

Symptoms

non-RAC - 112010(asm) - 112010(db) - 10205(db) - HP-UX Itanium (11.31)

After migrating 10.2 ASM to 11.2, now only the 10.2 database instances error out with ORA-29701: unable to connect to Cluster Manager


10.2 database alert log:
~~~~~~~~~~~~~~~~~~~~~~~~
ALTER DATABASE MOUNT
Fri Mar 11 14:26:17 2011
ORA-00202: control file:
'+NSHDEVHORA01_DATA/devphrm4/controlfile/current.394.683466541'
ORA-17503: ksfdopn:2 Failed to open file
+NSHDEVHORA01_DATA/devphrm4/controlfile/current.394.683466541
ORA-15001: diskgroup "NSHDEVHORA01_DATA" does not exist or is not mounted
ORA-15077: could not locate ASM instance serving a required diskgroup
ORA-29701: unable to connect to Cluster Manager

10.2 database sqlplus session:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
SQL> create tablespace test_asm datafile '+NSHDEVHORA01_DATA' size 10M;
create tablespace test_asm datafile '+NSHDEVHORA01_DATA' size 10M
*
ERROR at line 1:
ORA-01119: error in creating database file '+NSHDEVHORA01_DATA'
ORA-17502: ksfdcre:4 Failed to create file +NSHDEVHORA01_DATA
ORA-15001: diskgroup "NSHDEVHORA01_DATA" does not exist or is not mounted
ORA-15077: could not locate ASM instance serving a required diskgroup
ORA-29701: unable to connect to Cluster Manager


Additional information
~~~~~~~~~~~~~~~~~~~~~~~
* This is not affecting any of the 11.2 database instances or any non-ASM 10.2 database instances.
* Same OS software owner for all homes (11.2 asm_home, 11.2 or 10.2 db_home)
.

Cause


This issue looks similar to Bug 10054912 (DB IS UNABLE TO COMMUNICATE WITH ASM INSTANCE) and to do that in pre-11.2 and for HP only, we cannot use hostnames > 8 characters.

* Checked an 'uname' showing a 8 character value for the hostname. Also 'netconf' shows 8 characters for hostname.
* Checked the /tmp/.oracle socket files and the configuration suses a 12 character value for hostname.

In 11.1, there is a restriction for hostname, which is documented in 11.1 database release note on HP-UX IA:

http://download.oracle.com/docs/cd/B28359_01/relnotes.111/b32071/toc.htm#CHDEHEIB
6.5 Hostname Length Restriction for Oracle Clusterware on HP-UX Itanium
On HP-UX Itanium, the hostname for cluster member nodes must be no greater than 8 characters.
.

Solution

To fix this:

1) As root:

- set ORACLE HOME to GRID HOME:
- cd $ORACLE_HOME/crs/install
- perl roothas.pl -delete -force

2) As root:

Edit file: /etc/rc.config.d/netconf
- set the hostname and nodename to the first 8 characters ONLY
- reboot server

3) As root:

- set ORACLE HOME to GRID HOME:
- cd $ORACLE_HOME
- ./root.sh # this rebuilds the 'cluster' with the 8 character name.

4) As the Oracle Restart (gi_home) Software Owner :

Add/start all the components into the Oracle Restart configuration:

$GRID_HOME/bin/srvctl add listener -l LISTENER
$GRID_HOME/bin/srvctl start listener -l LISTENER

$GRID_HOME/bin/srvctl add asm -l LISTENER -d +ASM
$GRID_HOME/bin/srvctl start asm

$GRID_HOME/bin/srvctl add database -d -o -a
$GRID_HOME/bin/srvctl start database -d

5) Change to the 10g home and start up the 10g databases that use ASM, and it works.



來自 “ ITPUB部落格 ” ,連結:http://blog.itpub.net/17252115/viewspace-1298728/,如需轉載,請註明出處,否則將追究法律責任。

相關文章