From how to Configure or Re-configure Grid Infrastructure with Config.sh/config.bat (document ID 1354258.1)
Suitable for:
Oracle database-enterprise edition-version 11.2.0.2 and later
Information in this document applies to any platform.
Objective:
This article explains the config.sh (Config.bat on Windows) file under the $grid_home/crs/config/directory.
Details:
Basic:
"Config.sh" is a new feature of 11.2.0.2 Grid Infrastructure (GI). This feature is used for GI software to be installed or configure 11gR2 GI Cluster after cloning.
Config.sh will launch the GI configuration framework GUI, which will instruct the user to enter some necessary information through some pages, and finally config.sh will be prompted to run the root script. Config.sh can also be run in silent mode (with response files at this time)
Config.sh is not a tool for deploying software. Software can be deployed by running Oui or clonig.
This is to say: If the GI software is destroyed, config.sh will not help in this situation.
At this point (if the GI software is compromised), deinstall or node removal/addition procedure can be used in this case.
The case where config.sh can be used:
On all nodes, after using the rootcrs.pl script to deconfigured GI cluster
In this case, if the interactive mode, config.sh will ask the cluster parameter to generate GI configuration files and prompt to run the root.sh to establish a new GI cluster
Starting with config.sh asking for configuration information, config.sh won't care if the original cluster is a brand new installation or an upgrade from an older version,
There is no concern about whether the new cluster has the same number of nodes, or whether the configurations (OCR/VD location, network info etc) are the same as before.
For 11.2.0.3 or later versions, if DiskGroup attribute compatible.asm is set to 11.2.0.0 and above
Config.sh can be reused for OCR and VD (re-use) already exists diskgroups
From other cluster, the GI from clone.
This is part of the clone process, see:
Oracle Clusterware
Administration and Deployment Guide
11g Release 2 (11.2)
When GI just installed software
In this case:
---If you do not have an earlier Oracle Clusterware installation, CONFIG.SH will ask cluster parameter to generate a GI profile (configuration files) and prompt to run root.sh to configure a new GI Cluster
---If an oracle Clusterware installation already exists, CONFIG.SH will ask cluster parameter to generate a GI profile (configuration files)
and run rootupgrade.sh to upgrade a new GI cluster. In this case, there is no need to deconfigure anything before running config.sh.
---if there is already a GI standalone installed, config.sh will error, config.sh cannot be used to upgrade Oracle Restart
The case of config.sh cannot be used:
For GI cluster environment, CONFIG.SH will configure/reconfigure all nodes in cluster at all node downtime (down time).
As a result, config.sh is not the best choice for scenarios where there is no downtime to complete.
One or more nodes is have problem, but there was node or nodes that was running fine,
In this case, node removal/addition procedure can is used to avoid downtime.
One or more nodes be have problem, but there was node or nodes that was running fine, and the cluster is freshly insta lled without any patch set regardless how long it had been running-if patch set update (PSU) has been applied, which is fi NE, and cluster parameters is not changed since original configuration, EG:OCR/VD on same location, network Configurati On have not been changed etc, and grid_home is intact. Deconfig and reconfig on each problematic node can used (as root, execute "$GRID _home/crs/install/rootcrs . Pl-deconfig-force "Then" $GRID _home/root.sh ").
Parallel execution of Root script:
When establishing or upgrading a GI cluster with multiple nodes, you can run root script (root.sh or rootupgrade.sh) on the first node and wait for completion.
Then, execute root script (root.sh or rootupgrade.sh) on all other nodes (except for the last node) and wait for completion.
Finally, execute root script on last node (root.sh or rootupgrade.sh)
Known issue:
Bug 10158098-can not find ASM disks
This issue affects 11.2.0.2 only, refer to Document 1277148.1 for more details.
Bug 10055663-oui stuck after nodes list is specified (INS-42017 could be reported)
This issue affects 11.2.0.2 only, Patch 10055663 needs to be applied.
If INS-42017 still happens after the patch was applied, compare grid_home on all nodes to make sure they are the same, Espe cially $GRID _home/inventory/contentsxml/comps.xml; If there ' s a discrepancy, investigate for cause (different patches etc) and fix it.
Refer to note 1623645.1 If oracle_home_name are different on different nodes.
Bug 13409054-oui does not start with ' Exception in thread ' main ' Java.lang.UnsatisfiedLinkError:no Orainstaller in Java . Library.path "
This issue affects 11.2.0.3 on Solaris, it's fixed in 12.1 and the workaround are to modify <grid-home>/crs/config/co nfig.pl
Replace the following:
Sub Setldlibrarypath () {
My $osname =$^o;
$LD _library_path = $ORACLE _home. $dirSep. ' Lib '. $pathSep. $OUI _library_location;
$LD _library_path. = $pathSep. $ORACLE _home. $dirSep. ' Bin ';
$LD _library_path. = $pathSep. $ENV {Ld_library_path} if (exists ($ENV {ld_library_path}));
$ENV {Ld_library_path} = $LD _library_path;
}
With:
Sub Setldlibrarypath () {
My $osname =$^o;
$LD _library_path = $ORACLE _home. $dirSep. ' Lib '. $pathSep. $OUI _library_location;
$LD _library_path. = $pathSep. $ORACLE _home. $dirSep. ' Bin ';
$LD _library_path. = $pathSep. $ENV {Ld_library_path} if (exists ($ENV {ld_library_path}));
$ENV {Ld_library_path} = $LD _library_path;
My $libpath = $ENV {Ld_library_path};
$ENV {ld_library_path_64} = $libpath;
}
"Translated from MoS article" using Config.sh/config.bat to configure or re-configure GI
