4 node RAC and services

Hi All,
We currently run 5 databases on a 5 node RAC.. 11.2.0.2 GI and a mix of 11.2 and 10.2 databases. Obviously, services are defined for use by the application based on load on each node..
I am seeing something strange..
CLV1P_COMMON is a service created to run on CLV1P2 instance but during my tests over the weekend, the service was coming up on different node each time.. Why is this behavior ? Is this expected ? My understanding was that the service should run only on the node it services.. and in the case, where you specifiy PREFFERRED to more than one instance, it uses them randomly.. ( load balancing )
Can anyone guide me please ?
Thanks
/oracle/dba/output/STR1P/rman
camgmolpdb01 [CLV1P1]-> srvctl status service -d CLV1P
Service CLV1P1.aaaaaaaaaaaa.com is running on instance(s) CLV1P1
Service CLV1P2.aaaaaaaaaaaa.com is running on instance(s) CLV1P2
Service CLV1P3.aaaaaaaaaaaa.com is running on instance(s) CLV1P3
Service CLV1P4.aaaaaaaaaaaa.com is running on instance(s) CLV1P4
Service CLV1P_COMMON.aaaaaaaaaaaa.com is running on instance(s) CLV1P2
/oracle/dba/output/STR1P/rman
camgmolpdb01 [CLV1P1]-> srvctl config service -d CLV1P -s CLV1P_COMMON.aaaaaaaaaaaa.com
Service name: CLV1P_COMMON.aaaaaaaaaaaa.com
Service is enabled
Server pool: CLV1P_CLV1P_COMMON.aaaaaaaaaaaa.com
Cardinality: 1
Disconnect: false
Service role: PRIMARY
Management policy: AUTOMATIC
DTP transaction: false
AQ HA notifications: false
Failover type: NONE
Failover method: NONE
TAF failover retries: 0
TAF failover delay: 0
Connection Load Balancing Goal: LONG
Runtime Load Balancing Goal: NONE
TAF policy specification: NONE
Edition:
Preferred instances: CLV1P2
Available instances: CLV1P1,CLV1P3,CLV1P4
This is the command I used to create the service
srvctl add service -d CLV1P -s CLV1P_COMMON.aaaaaaaaaaaaaa.com -r "CLV1P2" -a "CLV1P1,CLV1P3,CLV1P4" -P BASIC -l primary -y AUTOMATIC -e SELECT -m BASICThanks in advance
Edited by: USER101 on Oct 15, 2012 1:59 PM
Edited by: USER101 on Oct 15, 2012 2:01 PM

Hi,
preferred and available is normally only evaluated during startup of the service.
So if you started you server in the order 1 then 2,3,4 as a result the service will get started on node 1 and not on the preferred node 2. Reason: Service is defined that it should start at least one instance. At the moment of startup there was only server 1 active, hence it is starting there. If node 2 (preferred) started after that, it is not evaluated, since Oracle does not want to move a service simply because of startup.
If then node 1 is rebooted, the service will get failed over. to the pereferred instance. If you then stop instance on node 2 again it will be started on a different node (not fix).
Important: The service does not fail to preferred instance, only if it is up.
As a workaround (if you need the service always to be started on node 2 if available and only run on another instance, when node 2 is not up) you can do this with a FAN Callout script.
Regards
Sebastian

Similar Messages

R12 Apps with 4 node RAC and 4 apps nodes - Post Install check list

Hi All,
We have a new Oracle Apps R12 instalaltion with 4 node RAC and 4 Apps nodes with load balancing and external web tier being done by an outside firm. We are assigned the responsibility of checking that everything has been configured properly both on the RAC side and Apps side. I haven't worked on RAC earlier. Please let me know what all need to be checked before approving the install done by the outside firm.
Thank you!

Please check below metalink notes
RAC Assurance Support Team: RAC and Oracle Clusterware Starter Kit and Best Practices (Generic) [ID 810394.1]
Using Oracle 10g Release 2 Real Application Clusters and Automatic Storage Management with Oracle E-Business Suite Release 12 [ID 388577.1]
Using Oracle 11g Release 2 Real Application Clusters with Oracle E-Business Suite Release 12 [ID 823587.1]
Oracle E-Business Suite and Oracle Real Application Clusters Documentation Roadmap [ID 745759.1]
745759.1 Oracle E-Business Suite and Oracle Real Application Clusters Documentation Roadmap
384248.1 Sharing The Application Tier file system in Oracle E-Business Suite Release 12
387859.1 Using AutoConfig to Manage System Configurations with Oracle E-Business Suite Release 12
406982.1 Cloning Oracle Applications Release 12 with Rapid Clone
240575.1 RAC on Linux Best Practices
265633.1 Automatic Storage Management Technical Best Practices
Loadbalancer
note 380489.1 Using Load-Balancers with Oracle E-Business Suite Release 12
note 727171.1 Implementing Load Balancing On Oracle E-Business Suite - Documentation For Specific Load Balancer Hardware
note 601694.1 How To Check Session Persistence On BigIP F5 And Cisco Ace Load Balancer Appliances
note 603325.1 Using Cisco ACE Series Application Control Engine with Oracle E-Business Suite Release 12
Installation R12
note 761564.1 Oracle Applications Installation and Upgrade Notes Release 12 (12.1.1) for Linux x86
note 402310.1 Oracle Applications Installation and Upgrade notes Release 12 (12.0) for Linux (32-bit)
note 559518.1 Cloning Oracle E-Business Suite Release 12 RAC-Enabled Systems with Rapid Clone
note 735276.1 Interoperability notes E-Business Suite R12 with Oracle Database 11gR1
Shared Applications Confiugurations:
note 380483.1 Oracle E-Business Suite Release 12 Additional Configuration and Deployment Options
note 384248.1 Sharing The Application Tier File System in Oracle E-Business Suite Release 12
Edited by: Amigo on Dec 20, 2010 9:58 AM

Single node file system to 3 node rac and asm migration

hi,
we have several utl_file and external table applications running on 10.2 single node veritas file system. and we want to migrate to 3 node RAC ASM environment. what is the best practices in order to succeed this migration during this migration. thanks.

1. Patch to 10.2.0.3 or 10.2.0.4 if not already there.
2. Dump Veritas from any future consideration.
3. Build and validate the new RAC environment and then plug in your data using transportable tablespaces.
Do not expect the first part of step 3 to work perfectly the first time if you do not have experience building RAC clusters.
This means have appropriate hardware in place for perfecting your skills.
Be sure, too, that you are not trying to do this with blade or 1U servers. You need a minimum of 2U servers to be able
to plug in sufficient hardware to have redundant paths to storage and for cache fusion and public access (a minimum of 6 ports).
And don't let any network admin try to convince you that they can virtualize the network paths: They can not do so successfully
for RAC.

Clone multi-node RAC and ASM to single node

Hi everyone.
I need clone system with 3 application server and 2 oracle database RAC and ASM to single-node. The operating system is RHEL 5.
I see some metalink notes, but we can't found nothing. I find notes with multi-node to single-node, but nothing with RAC to non-RAC.
The eBS version is 11.5.10.2 and database version is 10.2.0.3
Is possible this clone?
Thanks very much.

Hi User;
Please follow below and see its helpful:
EBS R12 with RAC and non-RAC
Re: RAC to single instance ebs R12
Re: Clone Oracle Apps 11.5.10.2 RacDB to Non-RAC DB
Re: CLONING R12 RAC to NON RAC CLONING giving error RMAN-05517 temporary file
[b]Migrating the DB-Tier (DB and CM) to Two node non RAC cluster[/b]
Also check:
http://www.oracle.com/technology/pub/articles/chan_sing2rac_install.html
Regard
Helios

Cluster with one 2 Node RAC and a Single Instance using ASM

Hi there,
i am not sure with one planned installation and want to ask, weather i am on the right track.
Some Facts:
Clusterware 11g
ASM 11g
Database 10gR2
AIX 5.3
3 Machines
2 Storages DS4700
My Plan
On Node 1 and Node 2 we install a RAC Database for an ERP Software
On Node 3 we install a single Instance Database for a Logistic Software
So i will install on all three Nodes Clusterware and an 3 Instances ASM - Cluster
I create 2 Diskgroups, one for the FRA and one for the Data, both on Luns on the DS4700
The RAC-Database and the Logistic-Database are using the same Diskgroups.
Is this the way to go for this circumstances?
The alternative is, as far as i see
Clusterware on an 3 Servers
One 2 Node ASM for the ERP Software
one Single Node ASM for the Logistcs
4 Diskgroups, because of the 2 ASM-Database 2 for the RAC and 2 for the Single Instance.
Please give me some hints, which way i should prefer.
My tendence is going to the first alternative. I like the idea to share the Diskgroups over more than on Database because of easy administration.
The load of the 2 Databases are completly different, the logistc software will nearly do nothing compared to the ERP Software, so this should'nt be a problem.
But maybe i oversee something, so please do not hesitate to tell me, i am completly wrong ;)
Thanks a lot
Jörg

Chris Slattery wrote:
why clusterware on 3rd machine ?
I'd have separate DGs but that's just me.If you wish to install ASM you need OCS installed on the machine, even if it is just one node at all.
It is a kind of a dependency, no OCS, no ASM
cu
Jörg

11gR2 dataguard using a primary 2 node RAC and a single standby

Hi,
I need to implement dataguard between 11gR2 RAC db to a single standby. How do I proceed?
Secondly, I need to timely apply the archive log i.e I need the dataguard to be in SYNC mode only in the night and the standby in read only
mode rest of the times. Can I achieve this using dataguard broker? If yes, can you please provide me the steps?
regards,
Ram

Doing that would require jobs again.
In your example:
9 am:
alter database recover managed standby database cancel;
-- here you can actually specify anything greater 780 because of the cancel at 10 pm; this is just to make sure the archive logs are being shipped.
alter database recover managed standby database delay=780 disconnect from session;
-- and, if your requriement is still to have the standby open read only
alter database open read only;
10 pm:
shutdown immediate;
startup nomount;
alter database mount standby database;
alter database recover managed standby database cancel;
alter database recover managed standby database using current logfile disconnect from session;
with that setup:
- your standby stops recovering at 9am
- the archive logs of the primary will still be shipped to your standby side for any emergency
- you'll have an open read only standby from 9am to 10pm
- the archive logs will be real time applied from 10pm to 9am
Note that this is just theory at the moment, I don't have a test set up like this. But it should give you an idea of what's possible and in what way you can accomplish your needs.

ORA-01652 unable to extend temp segment in 11gR2 with ASM and 2 node RAC

Hi ,
I have a 11.2.0.1 system with 2 node RAC and ASM . We are hitting ORA-01652 very frequently
and I can see space is available in TEMP .
CREATE TEMPORARY TABLESPACE "TEMP" TEMPFILE
SIZE 21474836480,
SIZE 137437904896
AUTOEXTEND ON NEXT 655360 MAXSIZE 131071M ( Will the next size ignored as the extent size is 1MB ?)
EXTENT MANAGEMENT LOCAL UNIFORM SIZE 1048576
And during the failure
SELECT tablespace_name,
total_blocks,
used_blocks,
free_blocks,
total_blocks*32768/1024/1024 as total_MB,
used_blocks*32768/1024/1024 as used_MB,
free_blocks*32768/1024/1024 as free_MB
FROM gv$sort_segment;
TABLESPACE_NAME TOTAL_BLOCKS USED_BLOCKS FREE_BLOCKS TOTAL_MB USED_MB FREE_MB
TEMP 4849568 2385728 2463840 151549 74554 76995
TEMP 0 0 0 0 0 0
Also when I checked from v$sort_usage I see only two session consuming space
SQL> SELECT b.tablespace tablespace,
2 -- b.segfile#,
3 -- b.segblk#,
4 b.blocks,
5 b.blocks * (SELECT value FROM v$parameter WHERE name = 'db_block_size')/1024/1024 bytes,
6 a.sid,
7 a.username username,
8 a.osuser osuser,
9 a.inst_id,a.sid,a.serial#,a.module
10 FROM gv$session a, gv$sort_usage b
11 WHERE a.saddr = b.session_addr and
12 b.blocks != 16 and
13 a.inst_id = b.inst_id
14 ORDER BY b.tablespace, b.blocks, bytes desc;
BLOCKS BYTES SID USERNAME OSUSER INST_ID SID SERIAL# MODULE
1,192,864 37277 381 CN obiqb 2 381 860 nqsserver@icmqapp01cdp (TNS V1
1,192,864 37277 3952 CN obiqb 2 3952 1374 nqsserver@icmqapp01cdp (TNS V1
2,385,728 74554 (mb)
Wat is causing the 1652 ?

877394 wrote:
Thanks Srini,
But I did try all that notes with no luck .
I have a temporary tablespace with 150gb (130+20) datafiles .
TABLESPACE_NAME FILE_NAME (BYTES/1024/1024)
TEMP +DATA_01/tempfile/temp.314.743964            131071
TEMP +DATA_01/tempfile/temp.312.757303             20480
and the DDL is
CREATE TEMPORARY TABLESPACE "TEMP" TEMPFILE
SIZE 21474836480,
SIZE 137437904896
AUTOEXTEND ON NEXT 655360 MAXSIZE 131071M
EXTENT MANAGEMENT LOCAL UNIFORM SIZE 1048576
And on querying v$sort_usage
SQL> SELECT b.tablespace tablespace,
2 -- b.segfile#,
3 -- b.segblk#,
4 b.blocks,
5 b.blocks * (SELECT value FROM v$parameter WHERE name = 'db_block_size')/1024/1024 bytes,
6 a.sid,
7 a.username username,
8 a.osuser osuser,
9 a.inst_id,a.sid,a.serial#,a.module
10 FROM gv$session a, gv$sort_usage b
11 WHERE a.saddr = b.session_addr and
12 b.blocks != 16 and
13 a.inst_id = b.inst_id
14 ORDER BY b.tablespace, b.blocks, bytes desc;
TABLESPACE BLOCKS BYTES SID OSUSER INST_ID SID SERIAL# MODULE
TEMP 63,328 1,979 3952 CN 2 3952 1374
TEMP 84,384 2,637 381 CN 2 381 860
TEMP 433,184 13,537 381 CN 2 381 860
TEMP 1,117,728 34,929 3952 CN 2 3952 1374
sum 1,698,624 53,082 (mb)
So the space utilization was less than 150gb and on querying v$sort_Segment
SQL> SELECT tablespace_name,
total_blocks,
used_blocks,
free_blocks,
total_blocks*32768/1024/1024 as total_MB,
used_blocks*32768/1024/1024 as used_MB,
free_blocks*32768/1024/1024 as free_MB
FROM gv$sort_segment;
TABLESPACE_NAME TOTAL_BLOCKS USED_BLOCKS FREE_BLOCKS TOTAL_MB USED_MB FREE_MB inst
TEMP 4849568 2311104 2538464 151549 72222 79327 2
TEMP 0 0 0 0 0 0 1
why for inst 1 it is showing 0 (which means no sort is happening in instance 1)
But still we got the error
ORA-1652: unable to extend temp segment by 32 in tablespace TEMP
01652, 00000, "unable to extend temp segment by %s in tablespace %s"
// *Cause: Failed to allocate an extent of the required number of blocks for
//          a temporary segment in the tablespace indicated.
// *Action: Use ALTER TABLESPACE ADD DATAFILE statement to add one or more
//          files to the tablespace indicated.

Can we create a (RAC) db service to assign a conc mgr to an instance

Hi Everyone,
Can we create a rac db service to for EBusiness? For example, can we create a service for AP module to run on instance1 of a two nodes RAC, and a service for PO to run on instance2.
I know i can assign a node to a conc mgr and use the specialization rule to include/exclude a user to send request to that conc mgr. But this seperation is done at the ebusiness, not at database service.
My understanding was that we can not create a service at the db because all the requests comes from one conc mgr. But having the ability to create more than one conc mgr, so maybe the question now is can i create a service for a conc mgr. lets say, I create a db service for standardard mgr on instance1 and create a db service for custom_mgr on node2. Is this possible?
I am not on the distribution list, please email me at [email protected]
Regards,
Manizha

Please see these docs/links:
How to run a concurrent program against a specific RAC instance with PCP/RAC setup? [ID 1129203.1]
PCP configuration
PCP configuration
Thanks,
Hussein

Dbms_schduler job is not running on a 2 node rac when 1st node fails

Hi,
I want to create a dbms_scheduler job in a 2 node RAC and the job should always run on the node1 and if node1 is down then it should run on node2. This is Oracle 10gR2 (10.2.0.3 in WINDOWS) .In order to do the same I did following
-- First Step
Using DBCA- Service Managment - Created a service (BATCH_SERVICE) and given node1 as preferred and node2 as available. This created following entry in tnsnames.ora in both nodes.
BATCH_SERVICE =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = node1-vip)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = node2-vip)(PORT = 1521))
(LOAD_BALANCE = yes)
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = BATCH_SERVICE)
(FAILOVER_MODE =
(TYPE = SELECT)
(METHOD = BASIC)
(RETRIES = 180)
(DELAY = 5)
--- Step 2
-- Created BATCH job classes.
BEGIN
DBMS_SCHEDULER.create_job_class(
job_class_name => 'BATCH_JOB_CLASS',
service => 'BATCH_SERVICE');
END;
-- Step 3 -- created a job using job_class as BATCH_JOB_CLASS
begin
dbms_scheduler.create_job(
job_name => 'oltp_job_test'
,job_type => 'STORED_PROCEDURE'
,job_action => 'schema1.P1'
,start_date => systimestamp at time zone 'US/Central'
,repeat_interval => 'FREQ=DAILY;BYHOUR=11;BYMINUTE=30;'
,job_class => 'BATCH_JOB_CLASS'
,enabled => TRUE
,comments => 'New Job.');
end;
Now when I monitor this job it runs on node1. Now I started testing for failover. I manually shutdown 1st instance. Then as per my understanding job should run on 2nd node. But job is not picking up.
when I run the followign command
srvctl status service -d db -s BATCH_SERVICE
service BATCH_SERVICE is running on instance node2.
Any help is really appreciated.

It does not show that whether job is running or broken.

2 Node RAC -Adminstrator Manged ORpolicy-managedDatabase - which is better

We are planning to build a 2-node RAC 11gR2 . I went through the documentation and found that administrator managed datbase is better when the no of nodes are very less .
Need your inputs on this as we are still in the designing stage of the setup.

Shashi,
It all goes back to your objective. As you mentioned, it is two node RAC and I assume that you are gonna use this cluster database for one application. If that is the case then go with Administrator managed database because then you will have more control and require less configuration. Policy managed is more useful when you are implementing a true grid solution where database is used as a service and multiple applications are accessing data via the database services provided inside the large real application cluster system. When I say database as a service, people may want to consolidate smaller databases into the bigger RAC system to reduce operating cost and overhead while providing same level of database access to the applications.
So to answer your question, go with Administrator managed RAC. I hope I have provided you enough information for your decision :-)
Regards
http://www.oraxperts.com

CONVERTING TWO NODE RAC ON WINDOWS TO SINGLE INSTANCE

I got question about a two node RAC on Windows.
One of our vendors is hosting our oracle enviroment and we are planning to bring in the datacenter in house and manage it ourselves.
Current they have a two node RACon windows
The new enviroment we are building here , we want to get rid of RAC and bring it back to a single instance environment.
How do i approach this?
1. My first option was to create a new database server and do the export dp from one of the node and import to the new single instance enviroment.
My conncern is can you do export from RAC database and import it to a non RAC enviroment?
2. My second option is to delete one node on the two node RAC and then do export dump or cold backup and copy the database files and necessary config files.
let me know how i can approach this.....

Okay guys after digging aroung i found the following approach.....
Cloning Oracle 10gR2 from RAC Environment to Single Instance Environment will raise the following error
ORA-29702: Error occurred in Cluster Group Service Operation.
The new instance is looking for CRS on new single instance non RAC environment.
For UNIX platform follow the following steps:
- Shutdown the instance which is cloned to non RAC.
- Shutdown all the listeners.
- Relink Oracle executable with rac_off option.
$ make -f ins_rdbms.mk rac_off
$ make -f ins_rdbms.mk ioracle
- Remove all the parameters from init.ora which points to second instance
and it is on other node.
- Remove cluster_database parameter
- Remove undo_tablespace parameter which points to second instance
For Windows Platform follow the following steps:
- Shutdown the instance which is cloned to non RAC.
- Shutdown all the listeners.
- Rename following file in %ORACLE_HOME%/bin directory: orarac10.dll to orarac10.dll.bak
Give me your thoughts, there is nothing on Metalink....

Database migration from non rac and non asm to RAC database with ASM

HI,
i have requirement to import one database ( db1) which running one node with out ASM and with out RAC , now i want to import that database with rman backup to another database ( db2) which running on different server with 3 node RAC and ASM ( disk group +DATA ) , can you please provide me the steps how i can migrate.? ( i have created db2 with dbca with RAC).
regards,
sri

Sri,
This could be done multiple ways. Simplest for you is to use DBCA:
http://download.oracle.com/docs/cd/B28359_01/install.111/b28264/cvrt2rac.htm

New implementation - Client having issues with RAC and Grid on Windows

Client is starting a new implmenetation on Windows 2008 R1, Oracle 11.1.0.7. They have a 3 node RAC and Grid Control. They are having a number of issues and support is working through those issues. They are asking if there are clients on RAC and Grid on Windows. The Wiki shows a few customers on Windows, but was hoping to get some feedback. Tks.

There are lots of customer applications that are based on Windows RAC. And as mentioned above, it is very well supported by Oracle.
However, because of the security vulnerabilities of Windows, I've observed that most of the financial institutions and listed companies opt for unix based servers for all high profile applications and run only low to medium risk application on Windows.
Regarding maintanence issues, I guess it's going to be pretty much the same; there shall always be lots of bugs, lots of patches and updates just like an yother software; doesn't really matter if you are on Unix or Windows systems. As long as you purchase support, it's all good.
It's pretty much same concern regarding Grid Control on windows.
Cheers.

Multi-Node RAC-EBS R12 document

Hi All,
I am looking for high level steps for a clone of a multi-node RAC-EBS install along with expected times.
Could someone please share such a document for R12?
Thanks for your time!
Regards,

Hi
It is mandatory to be certified for RAC clones.
Note: 783188.1 - Certified RAC Scenarios for E-Business Suite Cloning
You have not mentioned what kind of clone do you looking for whether it's Multi-Multi or Multi -Single,please see the below thread for multi to single instance
Please see this docs
Clone multi-node RAC and ASM to single node
Thanks
yakub

2 Node RAC with Single Node Standby Help

Hello Everyone -
I'm preparing to set up a 2 node RAC and use a single-node standby database.
I'm a bit confused on how to set up the Standby Redo Logs on the standby to use the LGWR functionality from the RAC instances.
Should I just create a set of standby redo logs (alter database add standby logfile...) like on a single-instance environment or do I need to create standby logfile threads (alter database add standby logfile thread. . . ) one for each cooresponding thread on the RAC?
It just seems odd that 2 RAC redo threads are able to write to one standby redo logfile at the same time (or maybe each RAC redo thread picks the first available standby redo log and writes to it??)
If anyone has some advice or insight, I'd appreciate it.
Andy

Hello Andy,
Did you get answer to your question about RAC with single standby?
Could you please share with me your experience if you have done so.
Thanks
Amit

4 node RAC and services

Similar Messages

Maybe you are looking for