Patching in RAC [message #304432] |
Wed, 05 March 2008 03:20 |
harshkumar
Messages: 34 Registered: February 2008 Location: Delhi
|
Member |
|
|
Hi,
I have to apply a bugfix patch in RAC environment. I go through readme file and according to my understanding i have to follow the below steps:
1. Shutdown instance on both nodes
2. upload the p5606847_10203_Solaris86-64.zip file on server and unzip
3. cd 5606847
4. opatch apply
I have the below queries:
1. Do i need to stop all the oracle services on both nodes ? or shutdown instances is enogh.
2. Whether the patch will automatically apply on both nodes or we need to patch individually all the nodes ? if yes, then any difference in the above steps.
Please provide your inputs.
Thanks
Harsh
|
|
|
Re: Patching in RAC [message #304678 is a reply to message #304432] |
Thu, 06 March 2008 01:46 |
harshkumar
Messages: 34 Registered: February 2008 Location: Delhi
|
Member |
|
|
Hi Michel,
Can you please confirm that whether the steps i have mentioned are correct or not.
I got the answers of my questions. Please confirm, if wrong:
We need to stop the all the services except cluster daemon. it will patch both the nodes.
Thanks
Harsh
|
|
|
|
|
|
|
Re: Patching in RAC [message #305100 is a reply to message #304432] |
Sat, 08 March 2008 08:25 |
harshkumar
Messages: 34 Registered: February 2008 Location: Delhi
|
Member |
|
|
Thank you all for the support ans suggestions.
Actually i have posted the query after reading the README file. But in this file they did not mentioned that whether we need to stop all the oracle services.
I have followed the same steps as i have mentioned in my earlier post and it patched the successfully.
Thanks
Harsh
|
|
|
|
Re: Patching in RAC [message #306377 is a reply to message #305106] |
Fri, 14 March 2008 00:00 |
ram anand
Messages: 244 Registered: February 2008 Location: india
|
Senior Member |
|
|
Hi,
Can u tell me a better book for the Oracle Real application Clustering, that book should contain all the need for installation and configuration of RAC in 10g and below mentioned topic should also covered
RHEL AS 4 Installation
Oracle 10g Installation with ASM & CRS (Automatic Storage Management, Cluster Ready Service)
Create Database Instance
Create a Physical Database
Create Datafiles and Tablespace
Installation of DB Scripts in the Physical database
Performance reports in single node
Additional Node Installation (Linux & Oracle 10g)
Creation of Database Instance in Additional Node
Adding the additional node to physical database
Adding of tablespace and data files in additional node
Reinstallation of DB Scripts in the Physical database
Performance reports in two nodes
Thanks,
|
|
|
|
|
Re: Patching in RAC [message #310223 is a reply to message #304432] |
Mon, 31 March 2008 10:58 |
divroro12
Messages: 105 Registered: March 2008 Location: Laurel, MD USA
|
Senior Member |
|
|
I will recommend u stop all oracle processes in both nodes before applying the patch.
Also make sure ssh is properly configured to assure propee communication b/w all nodes before patching.
However, do not stop the crs...
Take care...
divroro12
|
|
|
|
Re: Patching in RAC [message #337412 is a reply to message #310269] |
Wed, 30 July 2008 17:26 |
sunil_v_mishra
Messages: 506 Registered: March 2005
|
Senior Member |
|
|
when we apply patch we are touching the binaries of oracle.
Readme doc which is been provided by oracle have all the steps, every patch has its own step but while CPU patch,
use the correct opatch utility
set the opatch path if required.
.
Stop listener
stop instance
stop agent
put the db instance in blackout if you using grid, so that no alert should be generated
Take proper down time for applying patch.
It means Before doing opatch apply
Note :- By which command we have to give for applying patch is there in readme.txt
once patch is applied successfully please run the catcpu.sql if u applying cpu patch... and also ultrp.sql after starting the DB than start the listener.
We can apply patch simultaneously on both the node on rac but its always good to apply patch on one node so that if any error comes our one instance is up and running, business will not be impacted.
[Updated on: Wed, 30 July 2008 17:27] Report message to a moderator
|
|
|