Configuration of fault-tolerant NFS share storage for VmWare by using GLUSTERFS

We are using glusterfs more then one year. One of applications of this solution is NFS storage for VmWare.

Further, I will not describe process of configuration and process of configuration OS in details.

1. Install Linux Centos 6.3 x64 on the 300GB raid disk.
Format (ext3) and mount 600GB raid disk as /big. (use UUID identifier for /etc/fstab mount)
Make directory - /small

2. Configure network.
Create two bonds intefaces mode=6 ALB

echo "alias bond0 bonding" >> /etc/modprobe.d/bonding.conf 
echo "alias bond1 bonding" >> /etc/modprobe.d/bonding.conf

Configure interfaces (only one interface example)

cat /etc/sysconfig/network-scripts/ifcfg-bond0
DEVICE="bond0"
BOOTPROTO="static"
ONBOOT="yes"
IPADDR=172.18.0.1
NETMASK=255.255.255.0
TYPE="bond"
BONDING_OPTS="miimon=100 mode=6"

cat /etc/sysconfig/network-scripts/ifcfg-eth0 (for ifcfg-eth1 the same)
DEVICE="eth0"
BOOTPROTO="none"
ONBOOT="yes"
SLAVE="yes"
TYPE="ethernet"
MASTER="bond0"

3. Install software

cd /etc/yum.repos.d/
wget http://download.gluster.org/pub/gluster/glusterfs/3.3/3.3.1/CentOS/glusterfs-epel.repo

mkdir /backup
cd /backup
wget http://mirror.yandex.ru/epel/6/x86_64/epel-release-6-8.noarch.rpm
rpm -i epel-release-6-8.noarch.rpm

yum update
yum install mc htop blktrace ntp dstat nfs glusterfs* pacemaker corosync openssh-clients

4. Disable Selinux

cat /etc/selinux/config

# This file controls the state of SELinux on the system.
# SELINUX= can take one of these three values:
#     enforcing - SELinux security policy is enforced.
#     permissive - SELinux prints warnings instead of enforcing.
#     disabled - No SELinux policy is loaded.
SELINUX=disabled
# SELINUXTYPE= can take one of these two values:
#     targeted - Targeted processes are protected,
#     mls - Multi Level Security protection.

SELINUXTYPE=targeted

5. Flush iptables. (All accept)

6. Configure service to start at boot

chkconfig ntpd on
chkconfig glusterd on
chkconfig corosync on
chkconfig pacemaker on

6. Reboot servers

1. Configure hosts:

echo "172.18.0.1" gluster-vmware1" >> /etc/hosts
echo "172.18.0.1" gluster-vmware2" >> /etc/hosts

2. Configure ntpd

3. Configure GLUSTERFS (using server gluster-vmware1)

gluster peer probe gluster-vmware2
gluster volume create small replica 2 transport tcp gluster-vmware1:/small gluster-vmware2:/small
gluster volume create big replica 2 transport tcp gluster-vmware1:/big gluster-vmware2:/big
gluster start small
gluster start big

Confirm that the volume has started using the following command:

gluster volume info

Configure NFS access

gluster volume set small nfs.rpc-auth-allow <comma separated list of addresses or hostnames of vmware nodes to connect to the server>
gluster volume set big nfs.rpc-auth-allow <comma separated list of addresses or hostnames of vmware nodes to connect to the server>

4. Configure corosync

On both gluster servers

cat /etc/corosync/corosync.conf

compatibility: whitetank
totem {
      version: 2
      secauth: on
      threads: 0
      interface {
              ringnumber: 0
              bindnetaddr: 10.255.254.N # where N is IP address of front end network
              mcastaddr: 226.94.1.1
              mcastport: 5405
              ttl: 1
      }
}
logging {
      fileline: off
      to_stderr: no
      to_logfile: yes
      to_syslog: yes
      logfile: /var/log/cluster/corosync.log
      debug: off
      timestamp: on
      logger_subsys {
              subsys: AMF
              debug: off
      }
}
amf {
      mode: disabled
}

cat /etc/corosync/service.d/pcmk

service {
      # Load the Pacemaker Cluster Resource Manager
      name: pacemaker
      ver:  1
}
END

Create authkey on one server and copy it to another server.

corosync-keygen.

Restart services

service corosync restart
service pacemaker restart  

Check

crm status

Last updated: Fri Jan 11 09:31:55 2013
Last change: Tue Dec 11 14:33:11 2012 via crm_resource on gluster-vmware1
Stack: openais
Current DC: gluster-vmware1 - partition with quorum
Version: 1.1.7-6.el6-148fccfd5985c5590cc601123c6c16e966b85d14
2 Nodes configured, 2 expected votes
============
Online: [ gluster-vmware1 gluster-vmware2 ]

5. Configure pacemaker

crm
crm(live)#
crm(live)#configure
crm(live)configure#
crm(live)configure#crm configure property no-quorum-policy=true 
crm(live)configure#crm configure property stonith-enabled=false
crm(live)configure#primitive FirstIP ocf:heartbeat:IPaddr2 params ip="10.255.254.30" cidr_netmask="24" op monitor interval="30s" on_fail="standby"
crm(live)configure#commit

OS:

cat  /etc/sysctl.conf
# increase Linux TCP buffer limits
net.core.rmem_max = 8388608
net.core.wmem_max = 8388608
# increase default and maximum Linux TCP buffer sizes
net.ipv4.tcp_rmem = 4096 262144 8388608
net.ipv4.tcp_wmem = 4096 262144 8388608
# increase max backlog to avoid dropped packets
net.core.netdev_max_backlog=2500
net.ipv4.tcp_mem=8388608 8388608 8388608

GLUSTER:

gluster volume set big nfs.trusted-sync on
gluster volume set small nfs.trusted-sync on

1. Mount partitions to vmware nodes.

2. Creation of testing VM

We have deployed VM win7 x86 with thick disk 40GB. Also we have installed Performance Test on the VM

3. Results of tests

On experience with gluster I advise to use three node like more stable solution.
With this case don't forget to set for pacemaker:

crm configure property no-quorum-policy=false

for gluster:

gluster volume set VOLUMENAME cluster.quorum-count 2

Today i have updated the system to centos 6.4. Crm configuration command disappeared.
If you want to use crm configuration command with centos/redhat 6.4 you need to install crmsh: yum install crmsh
I'm using follow repo

[network_ha-clustering]
name=High Availability/Clustering server technologies (RedHat_RHEL-6)
type=rpm-md
baseurl=http://download.opensuse.org/repositories/network:/ha-clustering/RedHat_RHEL-6/
gpgcheck=1
gpgkey=http://download.opensuse.org/repositories/network:/ha-clustering/RedHat_RHEL-6/repodata/repomd.xml.key
enabled=1

Or you can use pcs (yum install pcs)

Profile of the author