Today during RAC installation I got the following error while running root.sh on the first node

[php]Setting the permissions on OCR backup directory
Setting up NS directories
PROT-1: Failed to initialize ocrconfig
Failed to upgrade Oracle Cluster Registry configuration[/php]

Then I checked ocrconfig log file and got the following output:

[php]Oracle Database 10g CRS Release 10.2.0.1.0 Production Copyright 1996, 2005 Oracle. All rights reserved.
2012-05-17 15:16:35.490: [ OCRCONF][1521206416]ocrconfig starts…
2012-05-17 15:16:35.492: [ OCRCONF][1521206416]Failure initializing ocr in DEFAULT. REBOOT INSTALL. err :[PROC-32: Cluster Ready Ser
vices on the local node is not running Messaging error [9]]
2012-05-17 15:16:35.492: [ OCROSD][1521206416]utstoragetype: /ocfs/ocrfile is on FS type 1952539503. Not supported.
2012-05-17 15:16:35.492: [ OCROSD][1521206416]utopen:6”: OCR location /ocfs/ocrfile configured is not valid storage type. Return c
ode [37].
2012-05-17 15:16:35.492: [ OCRRAW][1521206416]proprinit: Could not open raw device
2012-05-17 15:16:35.492: [ default][1521206416]a_init:7!: Backend init unsuccessful : [37]
2012-05-17 15:16:35.492: [ OCRCONF][1521206416]Failure in initializing ocr in INSTALL level. error:[PROC-37: Oracle Cluster Registry
does not support the storage type configured]
2012-05-17 15:16:35.492: [ OCRCONF][1521206416]Exiting [status=failed]…[/php]

There’re a lot of reasons and solutions for this error on the web, but in my case the problem was in the mount point parameter. When getting this error, the mount point parameter was as follows:

[php]/dev/dm-1 on /ocfs type ocfs2 (rw,_netdev,heartbeat=local)[/php]

Then I added “datavolume” parameter as follows:

[php]/dev/dm-2 on /ocfs2 type ocfs2 (rw,_netdev,datavolume,heartbeat=local)[/php]

And it worked!

Today, after changing password of the sysman user of the repository database of the Grid Control, I wasn’t able to connect to the Grid control and getting the following output:

503 Service Unavailable

This output means that there’s a problem with Repository database connection. In this case, you should check the OMS log files which reside under $OMS_HOME/sysman/log – emoms.trc

When I checked this file, I got the following error:

2012-05-10 15:50:37,740 [AJPRequestHandler-ApplicationServerThread-9] ERROR conn.ConnectionService verifyRepositoryEx.818 – Invalid Connection Pool. ERROR = ORA-01017: invalid username/password; logon denied

2012-05-10 15:51:08,004 [AJPRequestHandler-ApplicationServerThread-6] ERROR conn.ConnectionService verifyRepositoryEx.818 – Invalid Connection Pool. ERROR = ORA-01017: invalid username/password; logon denied

This means that OMS can’t connect to the repository database, as the password of the SYSMAN username was changed. In order to make OMS use new password, we need change two parameter in emoms.properties file which resides  $OMS_HOME/sysman/config directory:

oracle.sysman.eml.mntr.emdRepPwd=test

oracle.sysman.eml.mntr.emdRepPwdEncrypted=false

After making necessary changes I’ve tried to start OMS using the following command:

emctl start oms

But unfortunately I got the same error again … After working a while on this error, I’ve decided to take down all OMS applications using the following command and start them again:

$OMS_HOME/opmn/bin/opmnctl stopall

$OMS_HOME/opmn/bin/opmnctl startall

After all, while checking log file, the error disappeared.

If you’ve completely lost emoms.properties file, check the following metalink note to recreate it:

How to Recreate the emoms.properties File for 10g OMS from Scratch? [ID 733401.1]