CentOS 7.9 환경에서 DRBD + Pacemaker 고가용성 클러스터 구축 가이드

yum install -y epel-release elrepo-release
yum install -y drbd-dkms drbd-utils kmod-drbd

Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
 * elrepo: ftp.yz.yamagata-u.ac.jp
 * epel: d2lzkl7pfhq30w.cloudfront.net
No package drbd-dkms available.
Package drbd-utils-9.17.0-1.el7.x86_64 already installed and latest version
No package kmod-drbd available.

# 위와 같이 패키지 설치가 안될 경우 아래 방법으로 설치
yum install -y epel-release gcc make flex bison patch \
    kernel-devel-$(uname -r) elfutils-libelf-devel \
    openssl-devel xmlto asciidoc wget

cd /usr/src
wget https://linbit.com/downloads/drbd/9.0/drbd-9.0.29-1.tar.gz
tar -xzf drbd-9.0.29-1.tar.gz
cd drbd-9.0.29-1

make clean
make KDIR=/usr/src/kernels/$(uname -r)
make install

depmod -a
modprobe drbd

lsmod | grep drbd
modinfo drbd

resource r0 {
  device /dev/drbd0;   # DRBD 장치 이름 (가상 블록 디바이스)
  disk /dev/sdb1;    # 실제 물리 디스크 또는 파티션 경로
  meta-disk internal;
  on node1 {
    address 192.168.1.1:7789;
    node-id 0;
  }
  on node2 {
    address 192.168.1.2:7789;
    node-id 1;
  }
}

# 실제 물리 디스크는 반드시 umount & 초기화되어 있어야 합니다.
# 아래는 예시입니다.
umount /dev/sdb1
wipefs -a /dev/sdb1

drbdadm create-md r0
drbdadm up r0

# UP 확인 명령어
drbdsetup status

########################################
# 여기까지 Primary & Secondary 모두 수행 #
########################################

# Primary로 승격
drbdadm -- --overwrite-data-of-peer primary r0

# drbdadm -- --overwrite-data-of-peer primary r0 명령어가 안될 경우 아래 명령어 실행
# Primary로 강제 승격
drbdadm -- --force primary r0

mkfs.ext4 /dev/drbd0
mkdir /data
mount /dev/drbd0 /data

yum install -y pacemaker pcs corosync
systemctl enable pcsd --now
echo password | passwd hacluster --stdin
cp /etc/corosync/corosync.conf.example /etc/corosync/corosync.conf
systemctl enable --now pacemaker corosync

pcs cluster auth node1 node2
pcs cluster setup --name mycluster node1 node2
pcs cluster start --all
pcs cluster enable --all

# 오류 발생 시 Error: nodes availability check failed, use --force to override. WARNING: This will destroy existing cluster on the nodes. You should remove the nodes from their clusters instead to keep the clusters working properly.
# 모든 노드에서 클러스터 삭제
pcs cluster stop --all
pcs cluster destroy

pcs property set stonith-enabled=false
pcs resource defaults resource-stickiness=100
pcs resource create drbd_res ocf:linbit:drbd drbd_resource=r0 op monitor interval=15s
# Master(Primary)에서 진행
pcs resource master drbd_master drbd_res \
master-max=1 \ # 동시에 master가 될 수 있는 노드는 1개
master-node-max=1 \ # 한 노드에서 master가 될 수 있는 인스턴스는 최대 1개
clone-max=2 \ # 클러스터 전체에서 클론 인스턴스는 최대 2개 (노드 2개)
clone-node-max=1 \ # 한 노드당 1개만 클론 인스턴스 생성
notify=true # 상태 변화 시 notification (promote/demote) 동작 활성화

pcs resource create fs_res Filesystem device=/dev/drbd0 directory=/data fstype=ext4
pcs resource create vip_res IPaddr2 ip=192.168.1.100 cidr_netmask=24

pcs constraint order promote drbd_master then start fs_res
pcs constraint order start fs_res then start vip_res
pcs constraint colocation add fs_res with drbd_master INFINITY with-rsc-role=Master
pcs constraint colocation add vip_res with fs_res INFINITY

# 에러 발생 
pcs resource create drbd_res ocf:linbit:drbd drbd_resource=r0 op monitor interval=15s
Error: Agent 'ocf:linbit:drbd' is not installed or does not provide valid metadata: Metadata query for ocf:linbit:drbd failed: Input/output error, use --force to override

# 해결 방법
mkdir -p /usr/lib/ocf/resource.d/linbit/
wget -O /usr/lib/ocf/resource.d/linbit/drbd https://raw.githubusercontent.com/LINBIT/drbd-utils/master/scripts/drbd.ocf
chmod +x /usr/lib/ocf/resource.d/linbit/drbd
/usr/lib/ocf/resource.d/linbit/drbd meta-data #스크립트 정상 동작 확인 명령
다시 pcs resource create drbd_res ocf:linbit:drbd drbd_resource=r0 op monitor interval=15s

# 에러 발생
/usr/lib/ocf/resource.d/linbit/drbd: line 636:
/usr/lib/ocf/resource.d/linbit/drbd.shellfuncs.sh: No such file or directory

wget -O /usr/lib/ocf/resource.d/linbit/drbd.shellfuncs.sh https://raw.githubusercontent.com/LINBIT/drbd-utils/master/scripts/drbd.shellfuncs.sh
chmod +x /usr/lib/ocf/resource.d/linbit/drbd.shellfuncs.sh
# 다시 확인
/usr/lib/ocf/resource.d/linbit/drbd meta-data

yum install -y nfs-utils

echo "/data *(rw,no_root_squash,sync)" > /etc/exports
exportfs -r
systemctl restart nfs
systemctl enable nfs

# node-003
mount -t nfs [VIP]:/mount_target /mount_point
ex) mount -t nfs 192.168.0.100:/data /data

# node-001
reboot or init 0