Cluster file system

[satish]

Hi erman,

Could you please suggest on our below concerns

I want to mount the same disk to two linux machines and both machine must be able to read/write on the same disk at the same time.By the way,this requirement is to install EBS application to a shared disk.

I want to understand what i have to use in order to mount the same disk on two machines and how SAN or NAS are fitting in on all this...?

Is it required that cluster aware filesystem should be configured here in case of 2 node Linux?

Thanks,
Syed

[ErmanArslansOracleBlog]

If your purpose is to have a shared disk for Oacle EBS Apps Tier installation, then we are talking about a Shared Application Filesystem.

In order to build such an environment, Cluster filesystem is not a must. You can build such a configuration even with NFS..

In this types of configurations, I see mostly NFS is utilized rather OCFS or any other shared filesystem.

First, shared Appl_top using NFS is completely supported and documented.  The Oracle EBS system supports shared Application Filesystem terminology starting from application version 11i, and it’s supported until 12.2. It is easy to build and use.
Ofcourse, there are some points to be considered :

The sharing of APPL_TOP in a Windows environment is not supported.
INST_TOP also will be in shared locations unlike earlier releases.
All application tiers sharing the file system should be running on the same operating system.
Shared application tier file systems should be in read-write mode on all application tiers.

Please read : Sharing The Application Tier File System in Oracle E-Business Suite Release 12.2 (Doc ID 1375769.1)

Please also read : https://ermanarslan.blogspot.com/2017/08/ebs-122-important-point-to-be.html

[satish]

This is very informative blog.Thanks for posting it erman

https://ermanarslan.blogspot.com/2017/08/ebs-122-important-point-to-be.html

[satish]

Dear Erman,

When we request for reboot,linux team before they reboot the application server..they will be issuing the below command.Dont understand why they are issuing this command

systemctl stop pcsd

we are on 12.2.5
2node apps with a load balancer
2node db

I dont think,NFS is configured in our environment.Apps was installed in below mount points

/dev/mapper/prdapp_vgu02-prdapp_lvu02  100G   37G   64G  37% /u02--inst top
/dev/mapper/prdapp_vgu01-prdapp_lvu01  200G  145G   56G  73% /u01--appltop
/dev/mapper/prdapp_vgu03-prdapp_lvu03  200G   65G  136G  33% /u03--fs_ne


searched in fstab and i dont see the mount points in fstab

[applprod@app01inst]$ cat /etc/fstab

#
# /etc/fstab
# Created by anaconda on Tue Mar  7 16:07:11 2017
#
# Accessible filesystems, by reference, are maintained under '/dev/disk'
# See man pages fstab(5), findfs(8), mount(8) and/or blkid(8) for more info
#
/dev/mapper/rhel_prdapp01-root /                       xfs     defaults        0 0
UUID=5e4c2131-661d-400e-bd13-1e519ac7e4f8 /boot                   xfs     defaults        0 0
UUID=D537-179A          /boot/efi               vfat    umask=0077,shortname=winnt 0 0
/dev/mapper/rhel_prdapp01-tmp /tmp                    xfs     defaults        0 0
/dev/mapper/rhel_prdapp01-u04 /u05                    xfs     defaults        0 0
/dev/mapper/rhel_prdapp01-swap swap                    swap    defaults        0 0
/dev/mapper/rhel_prdapp01-swap00 swap                    swap    defaults        0 0
10.25.1.202:/vol/ERP    /ERPShareddrive         nfs     defaults,vers=3  0 0
[applprod@app01inst]$

Thank you

[ErmanArslansOracleBlog]

First of all, nfs mount must not be in fstab.
You can mount them with mount command as well.
Secondly,  you may have a 2 node EBS appiter without sharing the Application Filesystem.
I mean, every node can have its own filesystem, as well.

Okay.. If they use pcsd, then probably you have  a Redhat Linux Cluster.

But wait .. What is this? "10.25.1.202:/vol/ERP    /ERPShareddrive"

Anyways, it seems you have a cluster fs. (redhaat cluster based)
In this case, you should not add the file system to the /etc/fstab file because it will be managed as a Pacemaker cluster resource.

Ask this to your OS admin.. (ask the questions by reading by comments above)