|Oracle 11g Release 1 RAC On Linux Using VMware Server - This article describes the installation of Oracle 11g release 1 (11.1) RAC on Linux (Oracle Enterprise Linux 5) using VMware Server with no additional shared disk devices.|
which is the version of vmware server used?
is it possible to configure RAC between host and guest?
I used VMware Server 1.0.4 for this. That's the free version. The version is listed on the first VMware screen shot.
I don't think it is possible to this between the host and guest as the host can't see the shared storage and might not have two physical network cards.
sujesh said...Hi Tim,
in the VMware server installation you have written as "Regardless of the host OS, the setup of the virtual machines should be similar".
I was trying to install Vmware Server 1.0.4 over OEL5. but it asked lot of additional configuration questions and gave warning " Vmware server may crash because the gcc version is mismatch". I ignored this warning. after installation i tried to add OEL5 as guest os. but always boot failure. then i tried with OEL4 as host. then everything was fine. so VMware server 1.0.4 is not supporting OEL5 as host OS. please confirm this
after installation of oracle, how i can check the failover, load balancing etc in the system??
could you please give me some examples??
Questions in the forum please!
I run 1.0.4 on CentOS 5 with no problems. OEL5 should be fine.
For failover tests, start SQL*Plus on RAC2, but connect to RAC1 instance. Then turn off the power on RAC1, when you issue a query in your SQL*Plus session, you shold nw be connected to RAC2.
Questions in the forum!
Fernando said...The problems are with CentOS 5.1 and not with CentOS 5.0
It is hard to get CentOS 5.0 now...
James Radtke said...Amazing work here, especially to provide as "public knowledge"!
Do you accept paypal donations?
Montaser said...Very good job Tim as always.
david said...Note that de machines will has the same 'date'. if no has same date oracle show errors in the copy archives on other nodes.
OUI of CRS give you error with only 1seg on difers nodes
I have a problem with this command
The output from the RAC2 node is the same that the rac1 node
thanks for you help
If that were true then it means none of the configuration was done when you ran root.sh on the first node. Are you sure they are actually the same, or similar? The second one should have the message saying no keys were written.
PS. Questions in forum, not in comments please :)_
I tried to install clusterware but while running precheck with runcluvfy.sh it is showing error that the nodes are not clustered.I followed the exact procedure but it is asking password.
Can anyone help me to solve this.
Thanks & Regards,
Without wanting to brush you off, I have never seen this procedure fail if you have actually done what the article says with the same versions of the software the article uses.
The article you are following uses VMware Server 1.x, 32-bit OEL5 and 32-bit Oracle. There is an equivalent article for VMware Server 2.0, wth the same versions of OEL and Oracle.
If you are using 64-bit Linux and Oracle you are not doing the same thing and I cannot tell you if it will work or not.
With this in mind, are you really doing everything the article says? I find it hard to believe you are using VMware Server 1.x when version 2.x has been around for so long...
Tim... said...Oh yes. Asking for password implies to me that you've not set up SSH properly. Follow the article and test the password-less connectivity between the two hosts.
chulinn said...Hi Tom,
really helpful documentation, it helped me a lot.
I wanted to participate in the forum,
but the registration mail never arrives.
Could you please have a look? Thanks!
I've activated your account.
i have a problem with /dev/sdb1, /dev/sdc1, /dev/sd*1
When the OS reboot the devices comes with root:root permission. how could i fix them to stay in oracle:oinstall ... ?
The last part of creating the shared disks is to amend the "/etc/rc.local" file. This changes the owner and permissions of the devices.
Do what the article says and you will not have this problem.
Ram said...Hi Tim,
racgvip.log file having following error , Could you please help me out to figure it out what is the problem.
I have disabled the firwall on both guest as well as hosts.
2009-06-19 19:47:23.614: [ RACG] [ora.rac1.vip]: [color=#FF0000][b]ping to 192.168.2.1 via eth0 failed, rc = 1 (host=rac1.localdomain)
ping to 192.168.2.1 via eth0 failed, rc = 1 (host=rac1.localdomain)
Interface eth0 checked failed (host=rac1.localdomain)
2009-06-19 19:47:23.616: [ RACG] [ora.rac1.vip]: Invalid parameters, or failed to bring up VIP (host=rac1.localdomain)
2009-06-19 19:47:23.616: [ RACG] [ora.rac1.vip]: clsrcexecut: cmd = /u01/crs/oracle/product/11.1.0/crs/bin/racgeut -e _USR_ORA_DEBUG=0 54 /u01/crs/oracle/product/11.1.0/crs/bin/racgvip start rac1
2009-06-19 19:47:23.616: [ RACG] [ora.rac1.vip]: clsrcexecut: rc = 1, time = 10.000s
2009-06-19 19:47:27.383: [ RACG] [ora.rac1.vip]: clsrcexecut: cmd = /u01/crs/oracle/product/11.1.0/crs/bin/racgeut -e _USR_ORA_DEBUG=0 54 /u01/crs/oracle/product/11.1.0/crs/bin/racgvip check rac1
2009-06-19 19:47:27.383: [ RACG] [ora.rac1.vip]: clsrcexecut: rc = 1, time = 3.760s
2009-06-19 19:47:27.383: [ RACG] [ora.rac1.vip]: end for resource = ora.rac1.vip, action = start, status = 1, time = 15.470s
Deepak said...While installing the rpm, i got the errors like
" kernel-headers 2.2.1 is needed by glibc-headers ". Now my command of clusterverify
runcluvfy.sh stage -pre crsinst shows failed status because of the missing rpms.
In this sceneario, can i proceed with clusterware installation? If not, how to find the missing rpms?
It sounds like you didn't do the OS installation (with the correct package groups) that I specified. If you had, the kernel-headers package would already have been installed on your machine.
No, you can't proceed. You need to install the kernel headers or the software will not link properly. Install that package from your OS DVD.
Deepak said...Hi Tim,
I have re-installed the packages. I am using Oracle Enterprise Linux 5 Update 5.
I do NOT have following packages on Enterprise Linux 5 Disk 2
Therefore, runcluvfy.sh stage -pre crsinst shows failed.
Can you please tell me from where i can get these packages?
The location of the packages depends on the respin (update) of the OS. They have probably been moved onto another disk in your version. Just check the other discs.
Deepak said...Yes, you are absolutely right.
I do FIND the following packages on Enterprise Linux 5 Disk "3"
Raj said...I am getting following error while running rpm
[root@rac1 Server]# rpm -Uvh elfutils-libelf-devel-*
warning: elfutils-libelf-devel-0.137-3.el5.i386.rpm: Header V3 DSA signature: NOKEY, key ID 1e5e0159
error: Failed dependencies:
elfutils-libelf-i386 = 0.137-3.el5 is needed by elfutils-libelf-devel-0.137-3.el5.i386
libelf.so.1 is needed by elfutils-libelf-devel-0.137-3.el5.i386
febry said...cat /u01/crs/oracle/product/11.1.0/crs/cfgtoollogs/configToolFailedCommands
# Copyright (c) 1999, 2007, Oracle. All rights reserved.
/u01/crs/oracle/product/11.1.0/crs/install/onsconfig add_config rac1:6251 rac2:6251
/u01/crs/oracle/product/11.1.0/crs/bin/oifcfg setif -global eth0/192.168.2.0:public eth1/192.168.0.0:cluster_interconnect
plizzz help me !!!! already googling n no clue until now
The message you've sent gives no indication of the problem. Nothing I can comment on here.
If you want to pursue the issue, post a question on my forum, like the red text says. :)
Kartik said...Hi Tim ,
I am doing exactly the same thing with same vmware version , guest os being RHEL 4, the problem is one node does not see files in another node, though I am able to ping , ssh between nodes.Say I have put my database.zip in /opt/share11g in rac1 , however I am not able to see it from rac2- the cloned one of rac1. Could you please hint what exactly I may be doing wrong?
Sorry. No experience of doing this with RHEL4 and it's been years since I used VMware Server for any of these types of installations.
I don't know what shared disks you are talking about. The shared disks in my article are for use with ASM, not a normal file system.
PS. Please post questions in the forum.