Re: ORA-15003 on one node of RAC with ASM

I have had a look at that document and other bugs on MOS and I think none of them are applicable to our situation. We have only 3 diskgroups for this RAC cluster which are shared(concurrently mountable) by both nodes in the cluster.

After instance eviction, ASM on the evicted node refused to mount diskgroups and reported these errors.
When I tried to start ASM instance on the evicted node (without diskgroups mounted), Alert log of both ASM instances reported only themselves to be part of cluster. Nothing was reported in cssd/crsd logs about this action. In the OCRdump/olsnodes output both nodes were listed (as expected). 'cluvfy comp crs' for each local node reported successful results. Same cluvfy command if run clusterwide (-n node1,node2) reported the results to be unsuccessful.

Last night we performed CRS restart on the problematic node and then problem was gone.

Regards,

On Thu, Oct 24, 2013 at 8:43 AM, xueyuan.luo <xueyuan.luo_at_gmail.com> wrote:

> **
> Hi, Saurabh
> Do you have multiple diskgroups distributed on those 2 nodes?
> I suggest you review MOS Doc ORA-15003 DISKGROUP "DATA" ALREADY MOUNTED
> IN ANOTHER LOCK NAME SPACE (Doc ID 1531734.1).
> And see if this document can help you with find out the problem.
>
> Best Regards,
> Milo
>
>
> ERROR: diskgroup xxxx was not mounted
> ERROR: PST found another heartbeat (grp 2)
>
> On the console
> ORA-15032: not all alterations performed
> ORA-15003: diskgroup "xxxx" already mounted in another lock name space
>
> Looking at crs_stat -t output shows that each node can't see resources from
> another node. For example:
>
> On Node1:
> ora.node1.vip application ONLINE ONLINE node1
> ora....SM2.asm application ONLINE OFFLINE
>
> Node2:
> ora.node1.vip application ONLINE OFFLINE
> ora....SM2.asm application ONLINE ONLINE node2
>
>
>

--
http://www.freelists.org/webpage/oracle-l