Oracle VM RAC template - what it took

In my previous posting I introduced the latest Oracle Real Application Cluster / Oracle VM template. I mentioned how easy it is to deploy a complete Oracle RAC cluster with Oracle VM. In fact, you don't need any prior knowledge at all to get a complete production-ready setup going.

Here is an example... I built a 4 node RAC cluster, completely configured in just over 40 minutes - starting from import template into Oracle VM, create VMs to fully up and running Oracle RAC. And what was needed? 1 textfile with some hostnames and ip addresses and deploycluster.py.

The setup is a 4 node cluster where each VM has 8GB of RAM and 4 vCPUs. The shared ASM storage in this case is 100GB, 5 x 20GB volumes. The VM names are racovm.0-racovm.3. The deploycluster script starts the VMs, verifies the configuration and sends the database cluster configuration info through Oracle VM Manager to the 4 node VMs. Once the VMs are up and running, the first VM starts the actual Oracle RAC setup inside and talks to the 3 other VMs. I did not log into any VM until after everything was completed. In fact, I connected to the database remotely before logging in at all.

# ./deploycluster.py -u admin -H localhost --vms racovm.0,racovm.1,racovm.2,racovm.3 --netconfig ./netconfig.ini

Oracle RAC OneCommand (v1.1.0) for Oracle VM - deploy cluster - (c) 2011-2012 Oracle 
Corporation (com: 26700:v1.1.0, lib: 126247:v1.1.0, var: 1100:v1.1.0) - v2.4.3 - wopr8.wimmekes.net
(x86_64) Invoked as root at Sat Jun 2 17:31:29 2012 (size: 37500, mtime: Wed May 16 00:13:19 2012) Using: ./deploycluster.py -u admin -H localhost --vms racovm.0,racovm.1,racovm.2,racovm.3
--netconfig ./netconfig.ini INFO: Login password to Oracle VM Manager not supplied on command line or environment
(DEPLOYCLUSTER_MGR_PASSWORD), prompting... Password: INFO: Attempting to connect to Oracle VM Manager... INFO: Oracle VM Client (3.1.1.305) protocol (1.8) CONNECTED (tcp) to Oracle VM Manager (3.1.1.336) protocol (1.8) IP (192.168.1.40)
UUID (0004fb0000010000cbce8a3181569a3e) INFO: Inspecting /root/rac/deploycluster/netconfig.ini for number of nodes defined... INFO: Detected 4 nodes in: /root/rac/deploycluster/netconfig.ini INFO: Located a total of (4) VMs; 4 VMs with a simple name of: ['racovm.0', 'racovm.1', 'racovm.2', 'racovm.3'] INFO: Verifying all (4) VMs are in Running state INFO: VM with a simple name of "racovm.0" is in a Stopped state, attempting to start it...
OK. INFO: VM with a simple name of "racovm.1" is in a Stopped state, attempting to start it...
OK. INFO: VM with a simple name of "racovm.2" is in a Stopped state, attempting to start it...
OK. INFO: VM with a simple name of "racovm.3" is in a Stopped state, attempting to start it...
OK. INFO: Detected that all (4) VMs specified on command have (5) common shared disks between
them (ASM_MIN_DISKS=5) INFO: The (4) VMs passed basic sanity checks and in Running state, sending cluster details
as follows: netconfig.ini (Network setup): /root/rac/deploycluster/netconfig.ini buildcluster: yes INFO: Starting to send cluster details to all (4) VM(s)....... INFO: Sending to VM with a simple name of "racovm.0".... INFO: Sending to VM with a simple name of "racovm.1"..... INFO: Sending to VM with a simple name of "racovm.2"..... INFO: Sending to VM with a simple name of "racovm.3"...... INFO: Cluster details sent to (4) VMs... Check log (default location /u01/racovm/buildcluster.log) on build VM (racovm.0)... INFO: deploycluster.py completed successfully at 17:32:02 in 33.2 seconds (00m:33s) Logfile at: /root/rac/deploycluster/deploycluster2.log
my netconfig.ini
# Node specific information
NODE1=db11rac1
NODE1VIP=db11rac1-vip
NODE1PRIV=db11rac1-priv
NODE1IP=192.168.1.56
NODE1VIPIP=192.168.1.65
NODE1PRIVIP=192.168.2.2
NODE2=db11rac2
NODE2VIP=db11rac2-vip
NODE2PRIV=db11rac2-priv
NODE2IP=192.168.1.58
NODE2VIPIP=192.168.1.66
NODE2PRIVIP=192.168.2.3
NODE3=db11rac3
NODE3VIP=db11rac3-vip
NODE3PRIV=db11rac3-priv
NODE3IP=192.168.1.173
NODE3VIPIP=192.168.1.174
NODE3PRIVIP=192.168.2.4
NODE4=db11rac4
NODE4VIP=db11rac4-vip
NODE4PRIV=db11rac4-priv
NODE4IP=192.168.1.175
NODE4VIPIP=192.168.1.176
NODE4PRIVIP=192.168.2.5
# Common data
PUBADAP=eth0
PUBMASK=255.255.255.0
PUBGW=192.168.1.1
PRIVADAP=eth1
PRIVMASK=255.255.255.0
RACCLUSTERNAME=raccluster
DOMAINNAME=wimmekes.net
DNSIP=
# Device used to transfer network information to second node
# in interview mode
NETCONFIG_DEV=/dev/xvdc
# 11gR2 specific data
SCANNAME=db11vip
SCANIP=192.168.1.57

last few lines of the in-VM log file :

2012-06-02 14:01:40:[clusterstate:Time :db11rac1] Completed successfully in 2 seconds 
(0h:00m:02s) 2012-06-02 14:01:40:[buildcluster:Done :db11rac1] Build 11gR2 RAC Cluster 2012-06-02 14:01:40:[buildcluster:Time :db11rac1] Completed successfully in 1779 seconds
(0h:29m:39s)

From start_vm to completely configured : 29m:39s. The other 10m was the import template and create 4 VMs from template along with the shared storage configuration.

This consists of a complete Oracle 11gR2 RAC database with ASM, CRS and the RDBMS up and running on all 4 nodes. Simply connect and use. Production ready.

Oracle on Oracle.

Comments:

That is awesome. You mention "The setup is a 4 node cluster where each VM has 8GB of RAM and 4 vCPUs". The VM is on bare metal right? So that was 4 boxes with 8GB each? Just want to make sure I'm reading that right.

The shared storage was NAS? SAN?

Thanks,

LewisC

Posted by LewisC on June 04, 2012 at 10:04 AM PDT #

No, 1 physical server with 36GB RAM running Oracle VM. 4 VMs 8GB each running on that box. The storage in this case was direct attached but could be SAN or NAS, wouldn't matter.

Posted by Wim on June 04, 2012 at 10:13 AM PDT #

Thanks for the article . Is there any setup recommended for people who have only 8GB of server . How the shared storage can be presented to the vm guest being installed?Like is there any special setup to be done like any new filesystem or if we allocate everything under root the vmmanager can automatically find and allocate ?
Also the vm manager is on another server .

Posted by guest on June 04, 2012 at 11:08 AM PDT #

on an 8gb box.. well then you'd create 1.5-2gb vm's i guess. or you can create 3 node 2gb/vm clusters. if you have a second local disk you can use that for storage. the manager does run on a separate machine

Posted by wim on June 04, 2012 at 11:42 AM PDT #

Hi Wim,

I have a problem with using template with Oracle VM 3.1.
I can't find option to import template from /OVM/seed_pool - is there any workaround to import it without going through http(s) import ?

regards,
Marcin

Posted by Marcin Przepiorowski on June 22, 2012 at 12:45 PM PDT #

Too good.. Can I use this method to install on the physical servers also.. or is it limited to VM? There's a common SAN(FC) connectivity.

Posted by Vasu on June 29, 2012 at 01:20 AM PDT #

Post a Comment:
  • HTML Syntax: NOT allowed
About

Wim Coekaerts is the Senior Vice President of Linux and Virtualization Engineering for Oracle. He is responsible for Oracle's complete desktop to data center virtualization product line and the Oracle Linux support program.

You can follow him on Twitter at @wimcoekaerts

Search

Categories
Archives
« April 2014
SunMonTueWedThuFriSat
  
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
18
19
20
21
22
23
24
25
26
27
28
29
30
   
       
Today