Configuring OCFS2 Clustered File System on Debian Based Linux including Ubuntu and Kubuntu -

Configuring OCFS2 Clustered File System on Debian Based Linux including Ubuntu and Kubuntu

There's a lot of information and guides on OCFS2 for RHEL and Centos Linux but the package setup and configuration is slightly different and this has thrown some people off.

Installing OC2FS

You should install the following packages to get started:

apt-get install ocfs2-tools ocfs2console

Configure OC2FS

In RHEL/Centos the main configuration file is located in /etc/sysconfig/o2cb

However in Debian based Linux it is located in: /etc/default/o2cb (I don't think either make sense, why not just have a dir inside /etc/ called ocfs2 ?)

Open the file: /etc/default/o2cb and change the following line:

O2CB_ENABLED=false

to O2CB_ENABLED=true

This enables OCFS2 to startup at boot time (which you obviously should want if you are using such a filesystem).

Setup OC2FS cluster.conf

You'll need to manually create this path:

mkdir /etc/ocfs2

now touch /etc/ocfs2/cluster.conf

Here is the cluster.conf I created (you can have less/more nodes if you like):

node:
        ip_port = 7777
        ip_address = 192.168.1.115
        number = 1
        name = one
        cluster = ocfs2
node:
        ip_port = 7777
        ip_address = 192.168.1.116
        number = 2
        name = two
        cluster = ocfs2
cluster:
        node_count = 2
        name = ocfs2

The hostname must be correct for the node, meaning it has to be the actual hostname.

Under "cluster" that is the name you will be calling your cluster. Make sure that in /etc/default/o2cb the parameter matches (by default it is set to ocfs2 but it can be changed if you like).

Also make sure under each node that the cluster value is correct.  cluster is the name of the OCFS2 filesystem/cluster you want those nodes to belong to.

Move cluster.conf to each node (remember that each node must have the exact same cluster.conf)

Start the o2cb service: /etc/init.d/o2cb start

Loading module "configfs": OK
Mounting configfs filesystem at /sys/kernel/config: OK
Loading module "ocfs2_nodemanager": OK
Loading module "ocfs2_dlm": OK
Loading module "ocfs2_dlmfs": OK
Mounting ocfs2_dlmfs filesystem at /dlm: OK
Starting O2CB cluster ocfs2: OK

 Checking the status of o2cb should show something like this (if everything is correct):

/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 O2CB cluster ocfs2: Online
Heartbeat dead threshold = 7
  Network idle timeout: 10000
  Network keepalive delay: 5000
  Network reconnect delay: 2000
Checking O2CB heartbeat: Active

Format your OCFS2 partition:

mkfs.ocfs2 -b 4k -C 32K -L "MyOCFS2Cluster" -N 4 /dev/sda2

Make sure you replace "/dev/sda2" with the partition you want.  Don't make a mistake here because you'll lose all data on whatever partition you specify.

The "-N 4" stands for slots for up to 4 nodes, note that this can be increased but can't be decreased.  You shouldn't have a number higher than you need (or at least not much).  The Oracle documentation says you can always add more node slots though.

Mount your OCFS2 partition

mount /dev/sda2 /mnt/sda2

Debian/Kubuntu Error when mounting

On Kubuntu 8.04 I got this error, I'm not sure if the Debian package itself has this problem but here it goes:

mount.ocfs2: Node manager kernel module is not loaded probably because o2cb service not started

Everything is fine configuration file wise, but there is a bug in the init script provided by the Kubuntu Package.

You can fix it by: adding "load_module ocfs2_stackglue" to the LOAD_ACTIONS in /etc/init.d/o2cb

Or you can just manually modprobe ocfs2_stackglue

This is supposed to be fix in the latest package, but as of the time of this writing (April 7th, 2009) the latest package is the one I have.  I guess the udpated version hasn't been distributed yet.


  • PHP Migration from 5.3 to 5.4+ and dealing with deprecated functions
  • ffmpeg vidstab to stabilize video
  • userdel user userdel: cannot lock /etc/passwd; try again later.
  • mdadm how to mount inactive array
  • How to find and mount mdadm arrays automatically
  • M2Crypto.SSL.Checker.WrongHost: Peer certificate subjectAltName does not match host, expected fedora-archive.ip-connect.vn.ua, got DNS:mirror.ip-connect.vn.ua
  • [Wed Sep 20 15:34:44 2017] [notice] suEXEC mechanism enabled (wrapper: /usr/sbin/suexec) [Wed Sep 20 15:34:44 2017] [error] Init: Unable to read server certificate from file /www/ssl-certs/server.crt [Wed Sep 20 15:34:44 2017] [error] SSL Library Err
  • linux how to answer yes to copy
  • linux cp and mv will not overwrite due to alias!
  • ERROR 2006 (HY000) at line 567: MySQL server has gone away
  • vbulletin 4.2.5 after upgrading from 3.6 white screen fatal php errors
  • iptables v1.4.7: can't initialize iptables table `NAT': Table does not exist (do you need to insmod?) Perhaps iptables or your kernel needs to be upgraded.
  • Linux and FreeBSD how to set time and date
  • FreeBSD/OpenBSD OpenVPN Client error "Cannot allocate TUN/TAP dev dynamically"
  • kdenlive - No LADSPA plugins were found! Check your LADSPA_PATH environment variable. [producer_xml] failed to load transition "qtblend"
  • /usr/bin/supermin-helper exited with error status 1. To see full error messages you may need to enable debugging. See http://libguestfs.org/guestfs-faq.1.html#debugging-libguestfs at /usr/bin/virt-list-partitions line 177.
  • Linux Unable to mount cifs/smb share in /etc/fstab
  • MySQL: table is marked as crashed solultion
  • bash Linux how to get first or last letters of a word
  • l2tp ipsec VPN Error Sep 12 18:16:25 vps pluto[7299]: ERROR: asynchronous network error report on eth0 (sport=500) for message to 192.5.6.2 port 20640, complainant 192.5.6.2: Connection refused [errno 111, origin ICMP type 3 code 3 (not authenticated