cancel
Showing results for 
Search instead for 
Did you mean: 

Oracle RAC On SLES 10

Kapil Jha
Honored Contributor

Oracle RAC On SLES 10

Hi Gurus,

I am stuck in RAC installtion on SUSE 9.
it has 2 nodes and the node on which I am mounting ocfs2 filesystem it has
Checking heartbeat: Active
and other has

Checking heartbeat: Not active

when I am mounting the file system on second node it says

mount.ocfs2: Transport endpoint is not connected while mounting /dev/sdb1 on /oravl06_RAC

the only issue I could spot is port 7777 is closed on one server and open on other.

My question---
is this really related to port setting or something else.

How to open a port in SUSE?
whats should be the entry in services file for this port

7777/tcp open XXXX

BR,
Kapil+
I am in this small bowl, I wane see the real world......
8 REPLIES
Ivan Ferreira
Honored Contributor

Re: Oracle RAC On SLES 10

A TCP port is open when a service is started. For OCFS2 you need to ensure that the o2cb service starts correctly. Maybe you are having a problem with the cluster interconnect, that is, the private network interface between the hosts.

You could check with:

/etc/init.d/o2cb status

All information you need is in:

/usr/share/doc/ocfs2-tools-*/README.O2CB

Maybe the path is different for documentation in SUSE.
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?
Kapil Jha
Honored Contributor

Re: Oracle RAC On SLES 10

Hey Ivan,

everything seems to be fine in status, except heartbeat

# /etc/init.d/o2cb status
Module "configfs": Loaded
Filesystem "configfs": Mounted
Module "ocfs2_nodemanager": Loaded
Module "ocfs2_dlm": Loaded
Module "ocfs2_dlmfs": Loaded
Filesystem "ocfs2_dlmfs": Mounted
Checking cluster ocfs2: Online
Checking heartbeat: Not active

on the server where ocfs2 is not mounted.

I am not sure if its like this is this should be Active on both nodes.
I also feel somep problem with interconnect but could not figure out whatexactly is the scene.

BR,
Kapil
I am in this small bowl, I wane see the real world......
Ivan Ferreira
Honored Contributor

Re: Oracle RAC On SLES 10

Yes. It has to be active on both nodes:

Example:

node1>
Checking O2CB cluster sfx: Online
Heartbeat dead threshold: 31
Network idle timeout: 30000
Network keepalive delay: 2000
Network reconnect delay: 2000
Checking O2CB heartbeat: Active


node2>
Checking O2CB cluster sfx: Online
Heartbeat dead threshold: 31
Network idle timeout: 30000
Network keepalive delay: 2000
Network reconnect delay: 2000
Checking O2CB heartbeat: Active

Check your ocfs2/cluster.conf and identify the ip_address configuration on both nodes.

Ensure that you can ping that ip address from each one.

Use the ethtool command to identify if your private adapters have link.
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?
Kapil Jha
Honored Contributor

Re: Oracle RAC On SLES 10

All the links are UP I can ping from oe server to other.

The only issue I can feel is of port cause in one server port 7777 is open on other its closed.

What are the entries should be there in /etc/services file.

BR,
Kapil+
I am in this small bowl, I wane see the real world......
Ivan Ferreira
Honored Contributor

Re: Oracle RAC On SLES 10

I have no entries on /etc/services and everyhing is working correctly.

Any firewall running?
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?
Ivan Ferreira
Honored Contributor

Re: Oracle RAC On SLES 10

And also post the output of:

grep -i o2cb /var/log/messages*

On both nodes.
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?
Viktor Balogh
Honored Contributor

Re: Oracle RAC On SLES 10

# /etc/init.d/o2cb status
Module "configfs": Loaded
Filesystem "configfs": Mounted
Module "ocfs2_nodemanager": Loaded
Module "ocfs2_dlm": Loaded
Module "ocfs2_dlmfs": Loaded
Filesystem "ocfs2_dlmfs": Mounted
Checking cluster ocfs2: Online
Checking heartbeat: Not active

I would take a look into that init script and see what this "Checking heartbeat" does. My best bet is that it checks the port 7777, but I'm not sure as I don't know RAC environment in SuSE very well.
****
Unix operates with beer.
Ivan Ferreira
Honored Contributor

Re: Oracle RAC On SLES 10

I was digging a little more and the heartbeat is active when the OCFS2 filesystem is mounted.

That means that is inactive because you cannot mount the ocfs2 filesystem. We must find the reason of this.

Try with egrep "e2|ocfs2" /var/log/messages*

I have seen a problem similar searching and it was caused by different versions of ocfs2 installed on the nodes. Check the versions on both nodes with rpm -qa | grep ocfs
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?