Hi,

I successfully installed oracle 10g RAC on Sun solaris 10 on QFS mountpoint.

node1: having 2 local mountpoints "/oracle" and "/crsapp"
and
node2: having 2 local mountpoints "/oracle" and "/crsapp"

ocr, voting disks and database is on "QFS" file system.

Clusterware software is installed on "/crsapp" mountpoint and
oracle software is installed on "/oracle" mountpoint.

clusterware and oracle versions are 10.2.0.2.

"/oracle" and "/crsapp" mountpoints in both the nodes (node1 and nod2) are corrupted.
all the files in these mountpoints are deleted.

but i am having ocr and voting disks and datafiles on QFS, which are not corrupted.

How can I overcome this situation?
whether I need to install entire thing again?
before installing whether I need to delete or remove some files?
(for installation of CRS and oracle software)
how can I register the database which is already available in QFS mountpoint?

You do not have to go through the whoel installation process.
If you just want to recreate the ocr and voting disks in the same location, because they got corrupted (and you do not have a backup), you can simple rerun root.sh (after wiping ocr and voting with dd) from the nodes in serial order (one after the other). This will reformat the OCR and the voting disk.
After that continue as Chandra has described.

Message was edited by:
ssolbach

Hello,

while I agree with Chandra that the original problem was a corrupted Clusterware (DB) home and not the re-creation of corrupted Voting Disks or OCRs, I would like to remind us all that the re-run of the root.sh after a successfull installation is strongly discouraged. Please, refer to ML 220970.1 (Public RAC FAQ), where it says:

Is it supported to rerun root.sh from the Oracle Clusterware installation ?

Rerunning root.sh after the initial successful install of the Oracle Clusterware is expressly discouraged and unsupported. We strongly recommend not doing it.

In case where root.sh is failing to execute for the on an initial install (or a new node joining an existing cluster), it is OK to re-run root.sh after the cause of the failure is corrected (permissions, paths, etc.). In this case, please run rootdelete.sh to undo the local effects of root.sh before re-running root.sh.

Concluding, to recover from an entire loss of the OCRs and Voting Disks, please, take backups as described in the Oracle Clusterware Administration and Deployment guide.
If your Voting Disks and OCRs are corrupted after the initial install, you may want to find the root cause rather than just re-creating them.

Anyways, for the original problem, I would agree with Chandra's earlier post.

Thanks.

Hi,

Thanks all.

I have done follwoing things.

For CRS
1-CRS is cleaned using note 239998.1 10g RAC: How to Clean Up After a Failed CRS Install (to clean privous installation )
2-Install CRS
3-Apply patch set 10.2.0.2

For RDMS
1-Install RDBMS software only and run netca to create listener
2-Apply patch set 10.2.0.2

Reconfigure existing database
recreate init file,password file bdump,udump, etc
after this steps I can able to start instances using SQLPLUS from both nodes

Readd resources to CRS
Run srvctl add command to add database and instance back