Service creation error CRS-2664:Resource 'ora.prodcro.db' already running

Question

Hi,
Allow me to describe my architecture.
I have 2 2 node rac clusters - site 1 is the primary site and consists of 2 rac instances and site 2 is the standby site where a physical standby database is accessed by 2 rac instances.
Site 1 has services SERVICE1 and SERVICE2 configured and bound to the database instances PROD1 and PROD2.
When I try to created the same services on the standby site i can successfully create them using
srvctl add service -d SOCPRODCRO -s CROSERVICE1 -r PROD1 -a PROD2
srvctl add service -d SOCPRODCRO -s CROSERVICE1 -r PROD2 -a PROD1
However when i try to start them i get the following error
[oracle@mondbop303 {SOCPROD1} ~]$ srvctl start service -d SOCPRODCRO
CRS-2664: Resource 'ora.socprodcro.db' is already running on 'mondbop303'
CRS-2664: Resource 'ora.socprodcro.db' is already running on 'mondbop304'
CRS-2632: There are no more servers to try to place resource 'ora.socprodcro.service1.svc' on that would satisfy its placement policy
Can anyone tell me if they've seen this error before and give me any pointers on anything i'm doing wrong.
Cheers,
George

Answer 1

Hi Jorg,
Firstly thank you very much for taking the time and effort to reply to my post, I really appreciate it.
The Oracle version in use on my client’s site is 11g Release2 (11.2.0.1.0) running on Linux.
Also thank you very much for the pointers (and although i'm trying not to get my hopes up too much in case it doesn't fix the issue) i'm fairly confident that what you've suggested has hit the nail on the head and is the source of my problem i.e although you only explicitly say " on 10g you can create but not start a service on a standby side" it seems that this is also the case for 11g and that the issue lies with me attempting to start the service.
I was thinking that the fact i couldn't start the services on the physical standby site had to be something to do with the fact that it was a physical standby site and that therefore the database was only mounted.
Your pointer made me examine my session logs from last week and i did try to create the service as follows to automatically start when the Oracle Data Guard database role matches the physical_standby role but i got the same error e.g.
[oracle@mondbop303 {SOCPROD1} ~]$ srvctl add service -d SOCPRODCRO -s CROSERVICE1 -r SOCPROD1 -a SOCPROD2 -l physical_standby
[oracle@mondbop303 {SOCPROD1} ~]$ srvctl start service -d SOCPRODCRO
CRS-2664: Resource 'ora.socprodcro.db' is already running on 'mondbop303'
CRS-2664: Resource 'ora.socprodcro.db' is already running on 'mondbop304'
CRS-2632: There are no more servers to try to place resource 'ora.socprodcro.croservice1.svc' on that would satisfy its placement policy
However having read the srvctl documentation again for the add service section you directed me too i.e
=============================================================
"-l {[PRIMARY] |[PHYSICAL_STANDBY] |[LOGICAL_STANDBY] |[SNAPSHOT_STANDBY]}
The service role.
You use this option to indicate that the service should should only be automatically
started when the Oracle Data Guard database role matches one of the specified service roles."
=============================================================
my understanding from this documentation extract is that in fact i should actually be defining the
service i need on the standby cluster using the -l PRIMARY flag and that the service should only
be defined and then it would get automatically started when my standby database transitions to the
PRIMARY role during a switchover or failover. Is that your understanding Jorg?
The only way i can test this obviously is to define the services on my standby cluster, switchover to it, and see if the services are started against the former standby, now primary cluster database and if i can connect to the database using them once the database has failed over.
I'll be back on my customer site on Tuesday and i'll try this then and update the thread to let you know how i get on.
You also asked for more information about my environment:
My node names and (instance names) are as follows
active primary cluster is called cluster1
node names
mondbop001 (Oracle instance SOCPROD1)
mondbop002 (Oracle instance SOCPROD2)
Physical standby cluster is called cluster2
node names
mondbop303 (Oracle instance SOCPROD1)
mondbop304 (Oracle instance SOCPROD2)
Many thanks again.
Regards,
George