I am installing Oracle Grid Infratructure 126.96.36.199 on OEL 6.3
I reached the final step, running the root.sh on node 1. It succsessfully started the services but then I got this error:
Failed to create voting files on disk group OCR_VOTING. Change to configuration failed, but was successfully rolled back. CRS-4000: Command Replace failed, or completed with errors. Voting file add failed Failed to add voting disks at /opt/oracle/11.2.0/grid/crs/install/crsconfig_lib.pm line 6780.
+/opt/oracle/11.2.0/grid/perl/bin/perl -I/opt/oracle/11.2.0/grid/perl/lib -I/opt/oracle/11.2.0/grid/crs/install /opt/oracle/11.2.0/grid/crs/install/rootcrs.pl execution failed+
OCR_VOTING Diskgroup is configured using oracleasm:
oracleasm createdisk OCR_VOTING /dev/sdb1
/dev/sdb1 is a partition and has been created ot the LUN /dev/sdb using fdisk.
Any idea what might be wrong? Any answers are highly appreciated. Thanks!
Here is part of logfile of the root.sh - tool. Meanwhile I tried deinstalling the clusterware and cleaning up, then I verified the server configuration with cluvfy and all was OK.
+2013-01-04 00:24:26: Start of resource "ora.crsd" Succeeded+
+2013-01-04 00:24:26: Creating voting files+
2013-01-04 00:24:26: Creating voting files in ASM diskgroup OCR_VOTING
2013-01-04 00:24:26: Executing crsctl replace votedisk '+OCR_VOTING'
2013-01-04 00:24:26: Executing /opt/oracle/11.2.0/grid/bin/crsctl replace votedisk '+OCR_VOTING'
2013-01-04 00:24:26: Executing cmd: /opt/oracle/11.2.0/grid/bin/crsctl replace votedisk '+OCR_VOTING'
+2013-01-04 00:24:26: Command output:+ Failed to create voting files on disk group OCR_VOTING. Change to configuration failed, but was successfully rolled back. CRS-4000: Command Replace failed, or completed with errors. End Command output
+2013-01-04 00:24:26: Voting file add failed+
If you followed all of the prerequisites, you should not have a problem here...
1) DO NOT USE asmlib. IMProfessionalO it is superfluous.
2) make sure you are using more than one disk in a diskgroup
3) make sure the ownership and permissions are set correctly for the device(s)
4) you should use a minimum of 3 devices (1-2G should be sufficient) and use NORMAL redundancy for the OCR/Voting diskgroup. This will configure 3 voting files - one on each device.
5) configure udev or multipathd to properly set the ownership/permissions on reboot
6) make sure both the grid owner/group and the oracle owner/group can both READ and WRITE to ALL devices on ALL nodes BEFORE you start.
7) You should use the entire device (minus the first 1Mb - ie start at cylinder 2 to last)
To deconfigure what you have already done search for roothas.pl and then search the docs for the commands to "clean up" this failed install.
I have some more questions in connection with your answer. I followed the manual on http://www.oracle.com/technetwork/articles/hunter-rac11gr2-iscsi-088677.html which is using asmlib to configure the disks. Which other possibilities are there? I have been provided some LUN's from the network dept and this is what I have now.
Is it a best practice to use NORMAL redundancy or is it required. As the LUN's are placed on a filer, I used EXTERNAL redundany for OCR?
Obviously the cluvfy does not find such issues, as I had no errors reported before the installation.
Thanks in advance. I am still struggling with this problem, therefore your help is highly appreciated.