DRBD + Pacemaker & Corosync NFS Cluster Centos7

On Both Nodes

Host file
vim /etc/hosts

10.1.2.114 nfs1 nfs1.localdomain.com
10.1.2.115 nfs2 nfs2.localdomain.com

Corosync will not work if you add something like this: 127.0.0.1 nfs1 nfs2.localdomain.com - however you do not need to delete 127.0.0.1 localhost

Firewall

Option 1 Firewalld
systemctl start firewalld
systemctl enable firewalld
firewall-cmd --permanent --add-service=nfs
firewall-cmd --permanent --add-service=rpc-bind
firewall-cmd --permanent --add-service=mountd
firewall-cmd --permanent --add-service=high-availability

On NFS1

firewall-cmd --permanent --add-rich-rule='rule family="ipv4" source address="10.1.2.115" port port="7789" protocol="tcp" accept'
firewall-cmd --reloadfirewall-cmd --reload

On NFS2

firewall-cmd --permanent --add-rich-rule='rule family="ipv4" source address="10.1.2.114" port port="7789" protocol="tcp" accept'
firewall-cmd --reloadfirewall-cmd --reload
Disable SELINUX
vim /etc/sysconfig/selinux

SELINUX=disabled

Pacemaker Install

Install PaceMaker and Corosync

yum install -y pacemaker pcs

Authenticate as the hacluster user

echo "H@xorP@assWD" | passwd hacluster --stdin

Start and enable the service

systemctl start pcsd
systemctl enable pcsd

ON NFS1

Test and generate the Corosync configuration

pcs cluster auth nfs1 nfs2 -u hacluster -p H@xorP@assWD
pcs cluster setup --start --name mycluster nfs1 nfs2

ON BOTH NODES

Start the cluster

systemctl start corosync
systemctl enable corosync
pcs cluster start --all
pcs cluster enable --all

Verify Corosync installation

Master should have ID 1 and slave ID 2

corosync-cfgtool -s

ON NFS1

Create a new cluster configuration file

pcs cluster cib mycluster

Disable the Quorum & STONITH policies in your cluster configuration file

pcs -f /root/mycluster property set no-quorum-policy=ignore
pcs -f /root/mycluster property set stonith-enabled=false

Prevent the resource from failing back after recovery as it might increases downtime

pcs -f /root/mycluster resource defaults resource-stickiness=300
LVM partition setup

Both Nodes

Create a empty partition

fdisk /dev/sdb

Welcome to fdisk (util-linux 2.23.2).

Command (m for help): n
Partition type:
p primary (0 primary, 0 extended, 4 free)
e extended
Select (default p):(ENTER)
Partition number (1-4, default 1): (ENTER)
First sector (2048-16777215, default 2048): (ENTER)
Using default value 2048
Last sector, +sectors or +size{K,M,G} (2048-16777215, default 16777215): (ENTER)
Using default value 16777215
Partition 1 of type Linux and of size 8 GiB is set

Command (m for help): w
The partition table has been altered!

Create LVM partition

pvcreate /dev/sdb1
vgcreate vg00 /dev/sdb1
lvcreate -l 95%FREE -n drbd-r0 vg00

View LVM partition after creation

pvdisplay

Look in "/dev/mapper/" find the name of your LVM disk

ls /dev/mapper/

OUTPUT:

control vg00-drbd--r0

**You will use "vg00-drbd--r0" in the "drbd.conf" file in the below steps

DRBD Installation

Install the DRBD package

rpm --import https://www.elrepo.org/RPM-GPG-KEY-elrepo.org
rpm -Uvh http://www.elrepo.org/elrepo-release-7.0-3.el7.elrepo.noarch.rpm
yum install -y kmod-drbd84 drbd84-utils
modprobe drbd
echo drbd > /etc/modules-load.d/drbd.conf

Edit the DRBD config and add the to hosts it will be connecting to (NFS1 and NFS2)

vim /etc/drbd.conf

Delete all and replace for the following

include "drbd.d/global_common.conf";
include "drbd.d/*.res";

global {
usage-count no;
}
resource r0 {
protocol C;
startup {
degr-wfc-timeout 60;
outdated-wfc-timeout 30;
wfc-timeout 20;
}
disk {
on-io-error detach;
}
net {
cram-hmac-alg sha1;
shared-secret "Daveisc00l123313";
}
on nfs1.localdomain.com {
device /dev/drbd0;
disk /dev/mapper/vg00-drbd--r0;
address 10.1.2.114:7789;
meta-disk internal;
}
on nfs2.localdomain.com {
device /dev/drbd0;
disk /dev/mapper/vg00-drbd--r0;
address 10.1.2.115:7789;
meta-disk internal;
}
}

vim /etc/drbd.d/global_common.conf

Delete all and replace for the following

common {
        handlers {
        }
        startup {
        }
        options {
        }
        disk {
        }
        net {
                 after-sb-0pri discard-zero-changes;
                 after-sb-1pri discard-secondary; 
                 after-sb-2pri disconnect;
         }
}

On NFS1

Create the DRBD partition and assign it primary on NFS1

drbdadm create-md r0
drbdadm up r0
drbdadm primary r0 --force
drbdadm -- --overwrite-data-of-peer primary all
drbdadm outdate r0
mkfs.ext4 /dev/drbd0

On NFS2

Configure r0 and start DRBD on NFS2

drbdadm create-md r0
drbdadm up r0
drbdadm secondary all
Pacemaker cluster resources

On NFS1

Add resource r0 to the cluster resource

pcs -f /root/mycluster resource create r0 ocf:linbit:drbd drbd_resource=r0 op monitor interval=10s

Create an additional clone resource r0-clone to allow the resource to run on both nodes at the same time

pcs -f /root/mycluster resource master r0-clone r0 master-max=1 master-node-max=1 clone-max=2 clone-node-max=1 notify=true

Add DRBD filesystem resource

pcs -f /root/mycluster resource create drbd-fs Filesystem device="/dev/drbd0" directory="/data" fstype="ext4"

Filesystem resource will need to run on the same node as the r0-clone resource, since the pacemaker cluster services that runs on the same node depend on each other we need to assign an infinity score to the constraint:

pcs -f /root/mycluster constraint colocation add drbd-fs with r0-clone INFINITY with-rsc-role=Master

Add the Virtual IP resource 

pcs -f /root/mycluster resource create vip1 ocf:heartbeat:IPaddr2 ip=10.1.2.116 cidr_netmask=24 op monitor interval=10s

The VIP needs an active filesystem to be running, so we need to make sure the DRBD resource starts before the VIP

pcs -f /root/mycluster constraint colocation add vip1 with drbd-fs INFINITY
pcs -f /root/mycluster constraint order drbd-fs then vip1

Verify that the created resources are all there

pcs -f /root/mycluster resource show
pcs -f /root/mycluster constraint

And finally commit the changes

pcs cluster cib-push mycluster

On Both Nodes

Installing NFS

Install nfs-utils

yum install nfs-utils -y

Stop all services

systemctl stop nfs-lock &&  systemctl disable nfs-lock

Setup service

pcs -f /root/mycluster resource create nfsd nfsserver nfs_shared_infodir=/data/nfsinfo
pcs -f /root/mycluster resource create nfsroot exportfs clientspec="10.1.2.0/24" options=rw,sync,no_root_squash directory=/data fsid=0
pcs -f /root/mycluster constraint colocation add nfsd with vip1 INFINITY
pcs -f /root/mycluster constraint colocation add vip1 with nfsroot INFINITY
pcs -f /root/mycluster constraint order vip1 then nfsd
pcs -f /root/mycluster constraint order nfsd then nfsroot
pcs -f /root/mycluster constraint order promote r0-clone then start drbd-fs
pcs resource cleanup
pcs cluster cib-push mycluster

Test failover

pcs resource move drbd-fs nfs2

Other notes on DRBD

To update a resource after a commit

cibadmin --query > tmp.xml

Edit with vi tmp.xml or do a pcs -f tmp.xml %do your thing% 

cibadmin --replace --xml-file tmp.xml

Delete a resource

 pcs -f /root/mycluster resource delete db

Delete cluster

pcs cluster destroy
Recover a split brain

Secondary node
drbdadm secondary all
drbdadm disconnect all
drbdadm -- --discard-my-data connect all

Primary node
drbdadm primary all
drbdadm disconnect all
drbdadm connect all

On both
drbdadm status
cat /proc/drbd