quick and dirty VCS centos install howto


A short description on installing Veritas Cluster Suite (VCS) on CentOS 5 Linux. This will be installed separately from the Veritas Filesystem (VxFS) and Volume Manager (VxVM) to keep things small and easy.

I first used images from stacklet.com, but then switched to a plain CentOS 5.3 x64_64 install from CD. I only installed “base” in two virtualbox VMs.

Each had a 8GB root FS and 512MB of Ram. For an idling cluster + management, this has proven enough.

I also pre-build a router/service VM and multiple internal networks (cluster public, cluster heartbeat, cluster heartbeat) – I was too lazy to implement bonding interfaces 🙂

download veritas storage foundation ha 5.1 from Symantec
(to be honest I would suggest using sf 5.0 considering how much trouble I’ve had with the authentication stuff in 5.1)

download veritas cluster console from Symantec

download Centos 5 64bit

Build VMs:
2x Cluster Nodes,
1x Public Network (eth0)
2x Heartbeat Network (eth1,eth2)

install 64bit CentOS 5

edit /etc/redhat-release to give a version string from “real” redhat enterprise linux

install compat-libstdc++-33

set up dns forward/backward lookup (actually /etc/hosts will suffice)

set up ssh keys
ssh_keygen -t dsa on each node, get the public key from id_dsa.pub, add to authorized_keys on all nodes, use from=”fully qualified hostname” to keep safety to reasonable level

I disabled the local firewall using service iptables stop and chkconfig iptables off, but only because I don’t have the ports at hand. let me hope I get around to update this later.

extract the veritas installer

run installer

select preinstall check and list all the host names.

if it runs ok, do the install 🙂

Select “configure now”
Select “no global cluster”
And here is the status page with my other selections:

                  Veritas Cluster Server 5.1 Precheck Program
                              cent-vcs01 cent-vcs02

Cluster information verification:

        Cluster Name:      labclus
        Cluster ID Number: 2
        Private Heartbeat NICs for cent-vcs01:
                link1=eth1
                link2=eth2
        Low Priority Heartbeat NIC for cent-vcs01: link-lowpri=eth0
        Private Heartbeat NICs for cent-vcs02:
                link1=eth1
                link2=eth2
        Low Priority Heartbeat NIC for cent-vcs02: link-lowpri=eth0

Is this information correct? [y,n,q,b,?] (y)

And the summary for the admin virtual IP:


Cluster Virtual IP verification:

        NIC: eth0
        IP: 192.168.66.20
        NetMask: 255.255.255.0

Is this information correct? [y,n,q] (y)

The next question concerns secure mode, unfortunately I don’t have experience with the secure mode. The intro says it will use proper authentication and encrypt all cluster comms, so this is a definite “WANT!!!” feature for production and I’ll have to look into it more, but right now I can’t use it.

At the next question You can set up the admin username and password. The default “password” is just too dumb. Look, if you want to play unsafe, then just set it to blank, but I suggest you just use your damn brain and employ a real password.

After this I skipped over the smtp notifies and made the following settings for SNMP notifications:

SNMP notification verification:

        NIC: eth0
        SNMP port: 162
        Console: 192.168.66.1 receives SNMP traps for Warning or higher events

Is this information correct? [y,n,q] (y)

After some more OK’s it will just run and finish up the installation:

Logs are being written to /var/tmp/installer-201005020239sFN while installer
is in progress

    Starting VCS: 100%

    Estimated time remaining: 0:00                                    8 of 8

    Performing VCS configuration ...................................... Done
    Starting llt ...................................................... Done
    Starting gab ...................................................... Done
    Starting vxfen .................................................... Done
    Starting had ...................................................... Done
    Starting hashadow ................................................. Done
    Starting CmdServer ................................................ Done
    Performing VCS poststart tasks .................................... Done

Veritas Cluster Server Startup completed successfully

And finally here we go:

[root@cent-vcs01 rhel5_x86_64]# /opt/VRTSvcs/bin/hastatus -sum

-- SYSTEM STATE
-- System               State                Frozen

A  cent-vcs01           RUNNING              0
A  cent-vcs02           RUNNING              0

-- GROUP STATE
-- Group           System               Probed     AutoDisabled    State

B  ClusterService  cent-vcs01           Y          N               ONLINE
B  ClusterService  cent-vcs02           Y          N               OFFLINE

It WORKS! Just like that 🙂

p.s.:
and 3 hours after that I still didn’t get all the new great management framework up and running *OR* found a valid download url for the old java console. just a billion of annoying management and authentication daemons.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s