11gR2 Oracle Real Application Clusters RAC Build attempt, Toronto to Vancouver, 9-11 flight

http://www.masterschema.com for more information...

Interesting points: 1:55 Interrupted by flight attendant
6:38 All prereqs passed and install proceeds successfully
8:35 Oracle RAC seems to survive during turbulence
9:00 Conclusion: First attempt with RAC 11gR2 in VM a learning experience... and yes, I was tired.

On a flight segment from Toronto to Vancouver, September 11, 2009... I was bored during this flight and decided to record my first screencast of an experiment to install Oracle Real Application Clusters (RAC) 11gR2 (11.2.0.1) on 2 nodes in VMWare Fusion on a Mac OS X Snow Leopard box. A basic MacBook Pro rev 4,1, running 10.6.1 with 4gb of ram...

I was originally going to attempt this in the Y cabin (economy class) but they bumped me into J (First/Exec).

Before you watch the whole thing, please be aware I was NOT successful in completing the Grid Infrastructure portion of the RAC installation, but have since repaired the problem in my ntpd configuration and will upload a new screencast of a full RAC installation using VMWare.

Yes, I know the audio sucks. This was on a plane using the built-in microphone. This is also my first video...

I will put full procedures and tutorials up on http://www.masterschema.com when done

Category:

Education

Tags:

• Oracle
• RAC
• 11gR2
• Installation
• VMWare
• Fusion
• OS
• Snow
• Leopard
• 11g
• 11.2.0.1
• Real
• Application
• Clusters
• Enterprise
• Linux

[justiono]

Thanks for posting this. May I know what notebook spec/model do you use for this 11gR2 RAC?

[masterschema]

At 0:09, the screen shows Mac OS X 10.6.1, 2.4 Ghz Intel Core 2 Duo, 4GB 667 Mhz Ram.

I'm on a plane, so it's likely a notebook. =)

Was there a particular spec of interest?

[anilm2010]

Cool! vid :)! Thanks for the vid.

Can you do a Oracle 11g installation tutorial + backup and recovery for non-Oracle DBAs.

Cheers!

[masterschema]

I'll be working on some videos and written tutorials to help system administrators understand DBA work better...

and vice versa...

[antwoneeter]

@YoLninYo Yeah, but he's flying in uber first class - not really sitting next to anyone and you get to lay down.

[YoLninYo]

If someone was doing this shit next to me in a fucking flight I'd punch them in the fucking face. Imagine some weenie talking to his laptop for 3 hours straight. FFS!

[JUDOSAMURAIGB]

Are you using a shared disk in your virtual environment?

[melmak7]

Thnak a lot for your time and very goog tutorial.

The OUI finishes the installation unsucessfully in node2.

I think it´s for

Bug 9064614: CRSCTL CHECK CRS FAILS ON 2ND NODE 'VM-LNX-BTK29' CRS-5011

Do you had this problem before?

Thank you very much :)

[dikinuface]

thanx again. not knowing what exactly each command will do i have reservations about your procedure; based on the given process, it seems a bit extreme to drop and recreate a cluster just to get dbconsole up on a surviving node. is all this the only way to bring up EM/Grid Manager on a different node? it seems to me there would be an easier and quick way to do this as it seems like this loss of primary node senario might be a common occurence for people.

[bmaradi]

in a cluster environment the node from where you run emca -config dbcontrol db -cluster is the main node where the dbconsole is created.

you'll have to drop the repository:

ORACLE_HOME/bin/emca -deconfig dbcontrol db -repos drop -cluster

and then from one of the other remaining nodes:

ORACLE_HOME/bin/emca -config dbcontrol db -repos create -cluster

[dikinuface]

thanx. you said; "you need to recreate the dbconsole and pick one of the remaining nodes" ... how does one do this?

this is exactly what i cannot figure out how to do.

[bmaradi]

@dikinuface

the dbconsole for a 10gr2 rac environment is designed in this way:

on main node (the one where you executed emca in first place):

dbconsole (oc4j process)

emagent

on all other nodes:

emagent

so if the main node goes down then you need to recreate the dbconsole and pick one of the remaining nodes as the primary.

this was not the case for 10gr1 rac where each instance would have its own dbconsole

you can still revert to 10gr1 config for a 10gr2 rac dbconsole look it up in the otn docs

[dikinuface]

i have a 3 node RAC/GRID setup. EM runs from node 1 typically. if node 1 goes down, how can i run EM on one of the other nodes? the LISTENERS are up and dbconsole daemon starts but i cannot connect/start an EM session with either of the surviving nodes. only after i bring up node1 can i connect/run EM. what am i missing? is there a re-configure utility that tells oracle to allow EM to run on different nodes other than the original 1st node?

[anilm2010]

Thanks :) Looking forward to watching them