Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Mailing Lists -> Oracle-L -> Re: Further adventures in the world of RAC - Part 1

Re: Further adventures in the world of RAC - Part 1

From: K Gopalakrishnan <>
Date: Thu, 5 Apr 2007 10:04:05 -0700 (PDT)
Message-ID: <>


Can you query the /dev/oracleasm/disks and confirm whether it has the entries similar to the output from /etc/init.d/oracleasm listdisks. If you can see that , then manually startup the asminstance and before opening

issue alter system set asm_diskstring='ORCL:*','/dev/oracleasm/disks/' scope=both;

Then open the ASM instance. ALternatively you can change the asm_diskstring init.parameter to include the /dev/oracleasm/disks path before starting up.

Lte me know if this helps.  

Best Regards,
K Gopalakrishnan,

Author: Oracle Database 10g RAC Handbook, Oracle Press 2006

Co-Author: Oracle Wait Interface: Oracle Press 2004.


Thanks for the response. I believe I followed instructions correctly. As to cluvfy, I did run that but I agree with the comment that it is inaccurate. I was able to successfully install and patch CRS after which I ran cluvfy -stage predbsinst. The first time I ran it I ran it from the node on which I was doing the installation and it failed. It returned an error indicating that the permissions on the /tmp/CVI_10. differed on the two nodes in the cluster. I looked carefully at the files and also asked the system admin to look and other than timestamps we saw no differences. So for fun I ran cluvfy from the second node, it succeeded and said everything was fine and all tests passed. So for fun I went back and ran it again from the first node and this time it succeeded and said everything passed. I have seen this behavior with cluvfy in the past. On several occasions it has given inconsistent results.

The error I got when attempting to create the ASM instance was

ORA-12801: error signaled in parallel query server PZ99, instance banacek +ASM1 (1)
ORA-15186 ASMLIB error function [ASM_IO], error [3], mesg = [I/O Error]

It was at this point I decided that somehow I had not configured asmlib correctly (See my post Further adventures in the world of RAC - Part 2) decided to remove the ASM instance and that's where I got into the trouble noted in this post. I have used the information in the note you referenced but at this point the CRS install is not the problem, it is the ASM install that is causing difficulties.

I have not opened an SR yet, I will do that this morning but to be quite honest, on several occasions in the past when I have opened SRs regarding RAC problems and posted the answers to this list I have gotten the answers from this list much more quickly than I have from Oracle. It just doesn't seem like there is anyone from Oracle who is very experienced with RAC.

So, again, any thoughts or ideas are greatly appreciated.

Thank you.

Bill Wagman
Univ. of California at Davis
IET Campus Data Center
(530) 754-6208
-----Original Message-----

From: K Gopalakrishnan [] Sent: Wednesday, April 04, 2007 11:14 PM To: William Wagman; Subject: Re: Further adventures in the world of RAC - Part 1


I had quickly browsed thru your adventurees. :) Believe me RAC install is very simple and straight forward if you religiuoly complete the pre requisites. Lot of times we complete the RAC installs in less than few hours.

Couple of questions.

Did you run the clusterverify (cluvfy) sucessfully? Did you create the partitions for ASM disks and set the ownerships correclty.
Were you able to read the asm disks from all the nodes.? (simple dd) What is the EXACT error message from asm alert log?

In case if you have already a TAR/SR open let us know the SR# and some one in this list may be able to look the SR and help you.

If you want to cleanup the crs installs just have a look at the note Note:239998.1 Subject: 10g RAC: How to Clean Up After a Failed CRS Install
-- Received on Thu Apr 05 2007 - 12:04:05 CDT

Original text of this message