Re: 10204 RAC upgrade

  • From: Sanjay Mishra <smishra_97@xxxxxxxxx>
  • To: amit bansal <amit.bansal82@xxxxxxxxx>
  • Date: Mon, 15 Feb 2010 09:47:43 -0800 (PST)

Amit

As per the Rollong upgrade, Oracle has the following steps, Which look like 
that only Node1 is down in terms of CLuster respurce but it will copy all of 
the changes to other nodes and only need to take the other Nodes down when 
Nodes 1 is completed. Please correct if I am making wrong assumptions


8.9.2.1 Rolling Upgrade
By default, Oracle supports rolling upgrade for Oracle Clusterware. Complete 
the following steps only on one node:
        1. Stop isqlplus on the primary node by issuing the following command:
isqlplusctl stop

        2. Shut down all processes in the Oracle home on the node that might be 
accessing a database, for example Oracle Enterprise Manager Database Control:
$ emctl stop dbconsole
$ lsnrctl stop

        3. Shut down all services in the Oracle home on the node that might be 
accessing a database:
$ srvctl stop service -d db_name [-s service_name_list [-i inst_name]]

        4. Shut down all Oracle RAC instances on the node on which you intend 
to perform the rolling upgrade. To shut down Oracle RAC instances on individual 
nodes in the database, enter the following command where db_name is the name of 
the database:
$ srvctl stop instance -d db_name -i inst_name

        5. If an Automatic Storage Management instance is present, then shut 
down the Automatic Storage Management instance on the node on which you intend 
to perform the rolling upgrade. To shut down an Automatic Storage Management 
instance, enter the following command where node is the name of the node where 
the Automatic Storage Management instance is running:
$ srvctl stop asm -n node

        6. Stop all node applications on the node on which you intend to 
perform the rolling upgrade. To stop node applications running on a node, log 
in as the root user and enter the following command, where node is the name of 
the node where the applications are running:
# srvctl stop nodeapps -n node

        7. Perform the steps mentioned in Section 9.


Note:The following instructions are displayed on the Oracle Universal Installer 
screen. 
To complete the installation of this patch set, you must perform the following 
tasks on each node:
        1. Log in as the root user.
        2. As the root user, perform the following tasks:
        1. Shutdown the CRS daemons by issuing the following command:
CRS_home/bin/crsctl stop crs


        2. Run the shell script located at:
CRS_home/install/root102.sh
This script will automatically start the CRS daemons on the patched node upon 
completion.
        3. After completing this procedure, proceed to the next node and 
repeat. 

        8. To upgrade the rest of the nodes, go to the next node and perform 
Steps 2 to 6 and then log in as the root user and perform the following steps:
        1. Enter the following command to shut down the Oracle Clusterware:
# CRS_home/bin/crsctl stop crs


        2. Run the root102.sh script to automatically start the Oracle 
Clusterware on the patched node:
# CRS_home/install/root102.s

 
Sanjay




________________________________
From: amit bansal <amit.bansal82@xxxxxxxxx>
To: smishra_97@xxxxxxxxx
Cc: oracle-l@xxxxxxxxxxxxx
Sent: Mon, February 15, 2010 4:22:48 AM
Subject: Re: 10204 RAC upgrade

Hi Sanjay, 

You can do rolling patching for Clusterware. For Database and ASM, you can 
install 10.2.0.1 +10.2.0.4 patch in different home directly. Bring down 
database and start one instance and run catupgrd.sql

Then bring down db again and using srvctl change the ORACLE_HOME in OCR using 
srvctl command. You will also be required to copy the init*ora and password 
files in new Oracle home. I think this should fit into your 2-3 hrs window

But test this :)

Cheers
Amit
http://askdba.org/weblog/

On Mon, Feb 15, 2010 at 2:37 AM, Sanjay Mishra <smishra_97@xxxxxxxxx> wrote:

Hi
>
>I had to upgrade one of the 8 Node RAC environment and it is running 5 
>Database with ASM and RDBMS with it own Home in 10.2.0.3 on AIX. We have only 
>2-3 hrs window and so want to have expert opinion on the following
>       1. If I will upgrade CRS and it suggest at the end of successfull 
> install to run root102.sh on all NODES Can I run on all node in parallel. It 
> takes 10-15 minutes and so want to save some time. Not clear if it affect 
> each other as anyway CRS is down at that time on that node
>       2. Can I run the CRS upgrade indiviual on all node at the same time 
> from few new putty session instead of initiating from one window and 
> selecting all nodes. The reason is that it will take 15 Minutes to upgrade on 
> Node 1 and then 15 minutes to copy to other nodes. If there are 7 more nodes, 
> then it will take 15*7=100+ min. to copy to other nodes. So I f will run 
> seperately from each nodes, it will be total of 15-20 minutes for Software 
> and then 15 minutes for root102.sh.
>       3. If CRS is upgraded and is down, Can I start the ASM and RDBMS 
> upgrade at the same time or it will conflict due to /tmp/installer Lock files 
> or any other issue can come. 
>
>TIA
>Sanjay
>


      

Other related posts: