Upgrade from SLES12-SP5 pacemaker cluster with clvmd+LVM setup to SLES15-SP6

This document (000021505) is provided subject to the disclaimer at the end of this document.

Environment

SUSE Linux Enterprise High Availability Extension 15

Situation

Since SLE-HA 15, lvm2-lockd (lvmlockd) has replaced lvm2-clvm, in SLE-HA 15 SP2 lvm2-clvm has been removed.
This results in being unable to start cluster services after the upgrade due to the missing clvmd binary.
Additional steps are required both pre- and post-upgrade.

This how-to provides a general idea for the transition from clvmd to lvmlockd.
In more complex setups, the steps should be adjusted to match existing configuration.

Resolution

For simplicity, we will use a minimal example setup:

primitive clvmd clvm \
   op monitor timeout=90s interval=30s
primitive dlm ocf:pacemaker:controld \
   op monitor interval=60 timeout=60
primitive rsc-LVM LVM \
   params volgrpname=vg1 exclusive=true \
   op monitor interval=120s timeout=120s
primitive rsc-fs1 Filesystem \
   params device="/dev/vg1/lv1" directory="/fs1" fstype=xfs options="noatime,defaults" \
   op monitor interval=30s timeout=60s
primitive rsc-fs2 Filesystem \
   params device="/dev/vg1/lv2" directory="/fs2" fstype=xfs options="noatime,defaults" \
   op monitor interval=30s timeout=60s
group g-lvm-fs rsc-LVM rsc-fs1 rsc-fs2
group g-storage dlm clvmd
clone cl-storage g-storage \
   meta interleave=true ordered=true target-role=Started
order o-storage cl-storage g-lvm-fs

1. stop all resources accessing the clustered logical volumes

2. remove cluster flag:
vgchange -cn vg1

3. stop g-lvm-fs(reponsible for LV activation and mounting) and cl-storage(dlm+clvmd)
crm resource stop g-lvm-fs
crm resource stop cl-storage

4. remove clvmd and all LVM primitives
crm configure delete clvmd
crm configure delete rsc-LVM

5. stop and disable pacemaker on all nodes; choose any node and run
crm cluster run "systemctl disable --now pacemaker"

6. start the upgrade as described in our upgrade guide.

7. after successful upgrade of all nodes, update /etc/lvm/lvm.conf
In "global" section:
set: use_lvmlockd = 1
remove line: locking_type = 3

8. start pacemaker on all nodes
systemctl enable --now pacemaker

9. add new lvmlockd and LVM-activate primitives:
crm configure

primitive lvmlockd lvmlockd \
op start timeout="90" \
op stop timeout="100" \
op monitor interval="30" timeout="90"

primitive rsc-LVM LVM-activate \
params vgname=vg1 vg_access_mode=lvmlockd \
op start timeout=90s interval=0 \
op stop timeout=90s interval=0 \
op monitor interval=30s timeout=90s

modgroup g-storage add lvmlockd
modgroup g-lvm-fs add rsc-LVM before rsc-fs1
commit
quit

10. start cl-storage(dlm+lvmlockd)
crm resource start cl-storage

11. step #9 starts dlm and lvmlockd, which makes it possible to set lock-type to dlm:
vgchange --lock-type dlm vg1

12. start g-lvm-fs
crm resource start g-lvm-fs

Cause

Since SLE-HA 15, lvm2-lockd (lvmlockd) has replaced lvm2-clvm, SLE-HA 15 SP2 lvm2-clvm has been removed.

Additional Information

https://www.suse.com/releasenotes/x86_64/SLE-HA/15-SP2/index.html#jsc-SLE-9163
https://documentation.suse.com/sle-ha/15-SP6/single-html/SLE-HA-administration/#cha-ha-clvm

Disclaimer

This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.