Ask timesten 11.2 Using Oracle Clusterware to Manage Active Standby Pair

Using Oracle Clusterware to Manage Active Standby Pairs
http://download.oracle.com/docs/cd/E13085_01/doc/timesten.1121/e13072/cluster.htm#CCHCFAAD
What is that mean ?
I have to use Share Storage for... Oracle Clusterware?
I think right?... or wrong ?
Can I have VIP on Master node, and then master down... standby will active with VIP, right ?

A TimesTen A/S pair is a replicated configuation where a pair of TimesTen datastores, usually located on different machines, act in many ways as a single unit. The active master can process both queries and DML (insert/update/delete) while the standby master can only process queries (think of this as a little like Active Dataguard). Applications that run on the same machine as one of the TT datastores can connect to it via the very high performance 'direct mode' while applications running on other machines can access the datastores in client/server mode. There is a clearly defined procedure for failing over if the active master fails (which involves promoting the standby master to active) and for recovering a failed store back to the correct role. Direct mode applications must failover with the datastore whereas for client/server applications the connection(s) need to be failed over. The basic A/S pair replication mechanism has been available since TimesTen 6.0. The replication configuration is defined via special SQL syntax and monitoring and management (failover control etc.) is performed by way of built-in procedure calls. However, in TT 6.0 and 7.0 the actual monitoring and management of the A/S pair, including failover, is left for the user to implement themselvses or via some custom integration with an external cluster manager.
TimesTen 11g adds two major new capabilities:
1. Automatic client connection failover for client/server connections. Think of this as very similar to Oracle DB TAF and FAN. This does not require VIPs or Clusterware since it is implemented completely within TimesTen. However it works very well when used in onjunction with Clusterware.
2. A deep integration between TimesTen and Oracle Clusterware. All aspects of Timesten A/S pair definition, deployment, management and recovery are handled by Clusterware. There is just a single configuration file and a single TimesTen utility (ttCWAdmin) involved. Of course you need a properly configured Clusterware setup first which will require some form of shared storage (for OCR and voting disks) but TT storage (checkpoint and log files) does not need to be on shared storage. Setting up TT for use with Clusterware is very quick and easy (maybe 30 minutes the first time you do it and much quicker thereafter once you know what you are doing). From then on Clusterware will manage all aspects of failover and recovery completely automatically. VIPs are not required but can be used if desired e.g. for application failover purposes. Clusterware can also manage the failover of direct mode applications. Also, you can define automated backup cycles and spare nodes that can be used if a node suffers some permanent failure. The Clusterware integration offers very rich functionality but also very fast failover (typically just a few seconds in my testing).
Hope that helps clarify.
Chris

Similar Messages

Active/standby pair + Oracle db parameter FAILTHRESHOLD

Assume we have 3 databases
TT_A - timesten Active
TT_S - timesten Standby
O - oracle db.
SETURN MODE twosafe
Storage atributes
RETURN SERVICES OFF WHEN REPLICATION STOPPED;
FAILTHRESHOLD set to 10 value.
Two timesten databases are in consistent state.
Aplication update TT_A.
TT_S replication data to ORACLE.
Assume we stop replication.
So application can run on TT_A.
After 10 log switch TT_S will be marked as failed.
All logs waiting for TT_S will be delete.
So how oracle receiv the data ?

In active standby pair replication, cache operations are tightly coupled with replication. In normal operation of AWT cache group within an A/S pair, updates occur at the active which replicates them to the standby and then the standby pushes them to Oracle. The active and standby continually exchange housekeeping information about what transactions have been committed at the standby and which have been committed, via AWT, at Oracle. The active and the standby will only purge transaction logs for transactions that they know are safely stored in all 3 places. If the standby fails then, as long as you tell the active that it has failed (via a call to ttRepStateSave()), the active will take over the AWT push from the last transaction that it knows was safely committed in Oracle. No data will be lost.
If you are using oracle Clusterware to manage your A/S pair then you don't need to do anything as Clusterware will perform the necessary notification to the active that the standby has failed.
Chris

Can I use Oracle Clusterware 11 against 10.2.0.4 databases?

Implementing a 2 node RAC system on Linux RHEL 5, with version 10.2.0.4 databases. I read a fleeting post mentioning that you can use Oracle Clusterware version 11 against a 10R2 databases - is that the case? I've searched but can find no additional information regarding this, so if anyone can provide additional links or info, I would be most appreciative -
-kpw

Hi,
Implementing a 2 node RAC system on Linux RHEL 5, with version 10.2.0.4 databases. I read a fleeting post > mentioning that you can use Oracle Clusterware version 11 against a 10R2 databases - is that the case? Yes,
You can use the 11g clusteware to manage 10g and 11g RAC databases.
I've searched but can find no additional information regarding this, so if anyone can provide additional links or info, I would be most appreciative -
Take a look here [Understanding Compatibility in Oracle RAC Environments|http://download.oracle.com/docs/cd/E11882_01/rac.112/e10718/admin.htm#RACAD803], here [Oracle Clusterware Version Compatibility|http://download.oracle.com/docs/cd/E11882_01/rac.112/e10717/intro.htm#CWADD1118] and also in metalink note 337737.1 (Oracle Clusterware - ASM - Database Version Compatibility)
Hope it helps,
Cerreia
Edited by: Cerreia on 27/10/2009 20:38

Using Oracle Clusterware for keeping other things than RAC HA

Hello.
I'm wondering who else out there has used Oracle Clusterware to monitor items other than RAC.
There are some good examples on
http://www.oracle.com/technology/products/database/clusterware/index.html
But I was interested in finding other examples as well.
Thanks,
Erik

I am in the process of getting Clusterware to protect a single-instance (non-RAC) database... and not doing too well!
The theory behind it looks the same as for any non-Oracle app in that it just uses scripts (in this case Perl) to do the monitoring/stopping/starting.
My issue is once I've created the profile and registered it Clusterware just can't do a thing with it. It just errors and reports an unknown state. I have a SR on the go for it and it's been referred to the internal bods for review because everything looks right but it just doesn't work.
Has anybody achieved this and actually got it working?

Weblogic 6.1 Cluster using oracle clusterware

Hi,
I am very new to Weblogic Application Server. I have to set-up Weblogic Application Server 6.1 in cluster.The operating System is Solaris Sparc.
What I want to know is that is it possible to configure Weblogic Application Server in Cluster using Oracle Clusterware?
Also which cluterware will be a better solution for this Oracle Clusterware or Sun Clusterware?
If you can give me url of some doc regarding this also , I will be more than thankfull.
If anybody has any idea about it please reply it.
Thanks in advance
Vijay

Hi,
I am very new to Weblogic Application Server. I have to set-up Weblogic Application Server 6.1 in cluster.The operating System is Solaris Sparc.
What I want to know is that is it possible to configure Weblogic Application Server in Cluster using Oracle Clusterware?
Also which cluterware will be a better solution for this Oracle Clusterware or Sun Clusterware?
If you can give me url of some doc regarding this also , I will be more than thankfull.
If anybody has any idea about it please reply it.
Thanks in advance
Vijay

Updating RoutingServices using Oracle ESB Client Management API

Hi,
How to update Routing Rules of a RoutingService using Oracle ESB Client Management API.
Is there any other way other than console.
Any help is really appriciated.
Thanks

I decompiled HttpClient and ConsoleClientImpl and this is what I found --- The following URL is what is constructed,
http://localhost:80/esb/j_security_check and these parameters are passed as POST - j_username=oc4jadmin&j_password=welcome1
I tried
http://localhost:80/esb/j_security_check?j_username=oc4jadmin&j_password=welcome1
and got 404 error.
However, this exercise confirmed that hostname=localhost and port=80 is just fine.. the problem seems to lie somewhere else

Can I replicate new tables using the ACTIVE STANDBY PAIR replication scheme

Hi,
I have created myself a simple setup using an active/standby pair with a single subscriber like so:
CREATE ACTIVE STANDBY PAIR cie ON "tt-test1", cie ON "tt-test2" RETURN RECEIPT SUBSCRIBER cie on "tt-test3";
I have then added some tables on the master, they did not replicate automatically. I find this:
Command> repschemes;
Replication Scheme Active Standby:
Master Store: CIE on TT-TEST1
Master Store: CIE on TT-TEST2
Master Return Service: Return Receipt
Subscriber Store: CIE on TT-TEST3
Excluded Tables:
ROOT.EXTRACTOR_
ROOT.PROMPT_
ROOT.PREFERABLE_
Included Tables:
List too long (59 items), use verbosity 4 to display
<snip>
My question is ... how do I include these tables in replication?
Do I need to trash and clone the secondary master store and the subscriber again? Even doing that won't add the tables to the replication scheme so I don't think that is a solution.
I couldn't find much documentation on the ALTER REPLICATION statement but from what I could find it requires me to know the 'name' of the replication scheme and the examples in the documentation didn't work when I used 'Active Standby' as the scheme name in the statement.
Am I being retarded here? Is this a limitation of using the ACTIVE STANDBY PAIR replication model?
Thanks in advance.
Huw

When you setup and rollout the ACTIVE/STANDBY pair (or indeed legacy replication) it only includes tables that already exist. The normal deployment process is:
1. Create the first datastore (the one which will initially be the 'active').
2. Create (and populate) all necessary tables.
3. Create the active/standby pair replication scheme.
4. Start the repagent
5. Make the datastore active by calling ttRepStateSet('ACTIVE')
6. Use ttRepAdmin -duplicate to create the standby store from the active
7. Start repagent at standby
8. Use ttRepAdmin -duplicate to create the subscriber store from the standby
7. Start repagent at subscriber
If you need to add/remove tables later you must do the following:
At active node:
1. Create any new tables (and populate them) as needed
2. Stop repagent
3. Execute ALTER ACTIVE STANDBY PAIR with INCLUDE and/or EXCLUDE clauses as required
4. Start repagent
Then you need to redeploy the other stores:
At standby:
5. Stop repagent
6. Drop datastore (ttDestroy)
7. Re-create datastore from active using ttRepAdmin -duplicate
8. Start repagent
At subscriber:
9. Stop repagent
10. Drop datastore (ttDestroy)
11. Re-create datastore from standby using ttRepAdmin -duplicate
12. Start repagent
This is documented in the TimesTen Replication Guide in the section on administering an active/standby pair.
Chris

Anyone using VCS (Veritas Cluster Service) to manage Active-Standby

My requirement is to maintain an active-standby pair on primary site and two read-only subscribers at DR site.
We use VCS on our servers to manager all applications.
Our application requires immediate fail-over from active to standby in the event of a failure on active database.
I was told CRS is integrated with T10 and can take care of the automated fail-over.
Unfortunaltely, we have to keep VCS on these servers and wondering if any one has experiance on using VCS to manage the fail-over. Any scripts to share?

My requirement is to maintain an active-standby pair on primary site and two read-only subscribers at DR site.
We use VCS on our servers to manager all applications.
Our application requires immediate fail-over from active to standby in the event of a failure on active database.
I was told CRS is integrated with T10 and can take care of the automated fail-over.
Unfortunaltely, we have to keep VCS on these servers and wondering if any one has experiance on using VCS to manage the fail-over. Any scripts to share?

Can I use Oracle BPEL process manager with other AS and DB

Can we user oracle bpel process manager with jboss application server and postgresql database.. is it possible..
and what will be the price.. per licencse when i go to just purchase Oracle BPEL process manager

MAke sure you have not set ANY environment variable related to Oracle / BEA / Java / LD_library path. Use the following script to unset / set the initial settings:
#!/bin/sh
unset ORACLE_BASE ORACLE_HOME ORACLE_SID ORACLE_TERM
unset LD_LIBRARY_PATH LD_LIBRARY_PATH_64
unset CLASSPATH JAVA_HOME
export PATH=.:/usr/sbin:/usr/bin:/usr/local/bin:/opt/VRTS/bin
export BEA_HOME=/appl/oracle/products/9.2/weblogic
Marc
http://orasoa.blogspot.com

Error in HelloWorld service using oracle web services manager(Test Engine)

Hi All,
I deployed simple hello world web services on jwsdp2.0 and it is running at the following end point address :
http://localhost:8080/jaxrpc-HelloWorld/hello
When I use the Test engine provided by oracle web services manager to test this web service,Then I got the following response:
<Result><ServiceResult name="sayHelloBack"><Data>Could not send: JAXRPCSERVLET28: Missing port information</Data></ServiceResult></Result>
I am just wondering what is the cause of this behavior?Is it because of compatibility problem between oracle and jwsdp2.0 based web services?As I mentioned earlier that HelloWorld Service is running under jwsdp2.0 & tomcat50-jwsdp and it works fine from jwsdp2.0 based client.I would appreciate your reply.Thanks.
Kashif

I had the same problem, but adding the ant/lib directory to classpath helped.

WS-Security compatibility using Oracle Business Process Manager

Hi All,
Actually I like to invoke couple of secure services running on seperate server.These services are secured using WS-security.So I am just wondering how to configure Oracle Business process Manager so that process bpel file uses the certificates.I could not find any relevant documentation Is there any documentation available which explains how to do this?
I would appreciate your help.
Kind Regards
Kashif

yes we do native in Oracle BPEL PM (username token) - as explained here
http://www.oracle.com/technology/products/ias/bpel/documents/bpel_admin_10.1.3.1.0.pdf page 30
for more sophisticated ws sec stuff, pls use Oracle Webservice Manager

Anyone use Oracle Privileged Account Manager with OEM

Our security folks are pushing us to lock the "oracle" account and use OPAM. Have others done this? Is it possible?
What they are hoping we do is create a named account, and only su into oracle if need be. I realise these doesnt audit much of anything .. but alas..
If we lock "oracle" - can OEM pull passwords from OPAM? I think thats what they are looking for us to do. OPAM would rotate passwords and such. We as dbas should never need to know the oracle password.
I could be totally off on this ..
Anyone gone down this road before?
Daryl

Hi,
I want to check did you start the any service before running the security store? Go with create option as you are created different domain for OPAM.
Let me know.
Regards,
Ari

Securing Oracle SOA Applications by Using Oracle Web Services Manager

Guys!
I am following this tutorial since two days now.
http://www.oracle.com/technology/obe/fusion_middleware/owsm/secure%20soa/securing%20soa%20with%20owsm.htm
Everything goes fine till "Associating the Server Agent with the Web Service"
But when we go to section " Testing the Server Agent" My SOA suite doesnt let me do that. My OWSM and EM hangs and it doest let me enter.
Any idea what could be the problem?
Thanks
Deepak

Hi, I'm facing the same problem.
I've just installed the OWSM with the following options:
1. Oracle 10g Database
2. OC4J created on Oracle AS 10.1.2 (not standalone version OC4J)
3. Manual deploy to the OC4J and run "coresv initialize" successfully (I can query to the tables created with sqlplus)
Start the OC4J on the 10.1.2 AS and access to the login page with account "admin/oblix". It prompts "Login Failed".
Any idea on this problem? Is a "standalone OC4J" a must?
Thank you very much!
Message was edited by:
user517778

Using Oracle EM 11g with Active Data Guard 11g

Has anyone installed Active Data Guard 11g with OEM 11g for an active-active state or know if there are any installation procedures with these two products? We will be installing these products this year. Thanks, Cathy

If you already have both servers configured in Grid Control as targets, it should be fairly easy to create a standby database on server B. OEM has Availability tab on your primary database Instance section should have "Data Guard" section which has "Setup and Manage" option. You follow through the wizard, and it will generate a standby database on server B. Then you can convert your standby db on server B to Active Read-only database by clicking few more buttons.
If you do not have these two servers configured on your Grid Control, that would be the first task.
We set up three standby databases using Grid Control 11g in our test boxes, and they worked very well. Our biggest problem was to actually installing the Grid Control, but once installed, everything was easy.

Error during calling BPEL Process Using Oracle BPEL Process Manager Client

hi,
I have the following error during Calling BPel process and I can not detect the reason :
Exception in thread "main" java.lang.NoClassDefFoundError: javax/ejb/EJBException
     at com.oracle.bpel.client.util.ExceptionUtils.handleServerException(ExceptionUtils.java:76)
     at com.oracle.bpel.client.delivery.DeliveryService.getDeliveryBean(DeliveryService.java:254)
     at com.oracle.bpel.client.delivery.DeliveryService.request(DeliveryService.java:83)
     at com.oracle.bpel.client.delivery.DeliveryService.request(DeliveryService.java:53)
     at portlet.BPELProcess.startProcess(BPELProcess.java:68)
     at portlet.BPELProcess.main(BPELProcess.java:89)
My Code :
locator = new Locator("default", "welcome1");
deliveryService = (IDeliveryService)locator.lookupService(IDeliveryService.SERVICE_NAME);
String SchemaInputXML = "<soap:Envelope xmlns:soap=\"http://schemas.xmlsoap.org/soap/envelope/\">\n" +
" <soap:Header/>\n" +
" <soap:Body xmlns:ns1=\"http://www.globalcompany.com/ns/sales\">\n" +
" <ns1:User>\n" +
" <ns1:FirstName>787878</ns1:FirstName>\n" +
" <ns1:LastName>ujfyytytry</ns1:LastName>\n" +
" </ns1:User>\n" +
" </soap:Body>\n" +
"</soap:Envelope>\n";
NormalizedMessage nm = new NormalizedMessage();
nm.addPart("payload", SchemaInputXML);
deliveryService.request("userBPEL", "initiate", nm); --> here exception appear
userBPEL is my Process Name
initiate is BPel method

Message was edited by:
Marc Kelderman

Ask timesten 11.2 Using Oracle Clusterware to Manage Active Standby Pair

Similar Messages

Maybe you are looking for