How to create a promotable resource in Pacemaker

Promotable resources model an active/passive service where only one node holds the primary role while other nodes remain ready as secondaries. This fits replicated storage and databases that must enforce a single writer, while still allowing Pacemaker to promote a standby during failover.

In Pacemaker, promotable resources are managed as a Promotable Clone with per-instance roles such as Promoted and Unpromoted. The pcs CLI converts an existing base resource into a promotable clone (commonly named <resource_id>-clone) and exposes role-aware meta attributes for scheduling and constraints. The example below uses the ocf:pacemaker:Stateful agent to demonstrate the workflow.

Promotion and demotion switch roles at the cluster level; underlying replication and fencing must be correct to avoid dual-primary split-brain and data loss. Older stacks may use Master and Slave terminology and older meta attribute names, so version-specific naming differences must be accounted for when following examples.

Steps to create a promotable resource in Pacemaker:

Check cluster status before changing resource roles.

$ sudo pcs status
Cluster name: clustername
Cluster Summary:
  * Stack: corosync (Pacemaker is running)
  * Current DC: node-03 (version 2.1.6-6fdc9deea29) - partition with quorum
  * Last updated: Wed Dec 31 09:29:16 2025 on node-01
  * Last change:  Wed Dec 31 09:29:05 2025 by root via cibadmin on node-01
  * 3 nodes configured
  * 5 resource instances configured

Node List:
  * Online: [ node-01 node-02 node-03 ]
##### snipped #####

Confirm the base resource exists.

$ sudo pcs resource config stateful-demo
Resource: stateful-demo (class=ocf provider=pacemaker type=Stateful)
  Operations:
    demote: stateful-demo-demote-interval-0s
      interval=0s timeout=10s
    monitor: stateful-demo-monitor-interval-10s
      interval=10s timeout=20s role=Promoted
    monitor: stateful-demo-monitor-interval-11s
      interval=11s timeout=20s role=Unpromoted
    notify: stateful-demo-notify-interval-0s
      interval=0s timeout=5s
    promote: stateful-demo-promote-interval-0s
      interval=0s timeout=10s
    reload-agent: stateful-demo-reload-agent-interval-0s
      interval=0s timeout=10s
    start: stateful-demo-start-interval-0s
      interval=0s timeout=20s
    stop: stateful-demo-stop-interval-0s
      interval=0s timeout=20s

Replace stateful-demo with the resource ID intended for promotion.

Make the resource promotable.
```
$ sudo pcs resource promotable stateful-demo --wait=120
Resource 'stateful-demo-clone' is promoted on node node-03; unpromoted on nodes node-01, node-02.
```
The promotable clone resource ID is commonly created as stateful-demo-clone.

Use pcs resource master on older pcs releases.

Promotable resources on replicated storage require working fencing and safe replication, or promotion can cause split-brain and data loss.
Set promotable limits for the clone.
```
$ sudo pcs resource update stateful-demo-clone promoted-max=1 promoted-node-max=1 clone-max=2 clone-node-max=1
```
promoted-max limits promoted instances cluster-wide; promoted-node-max limits promoted instances per node; clone-max limits total instances; clone-node-max limits instances per node.

Use master-max and master-node-max on older releases.

Verify the promotable resource status.

$ sudo pcs status resources
  * Resource Group: web-stack:
    * cluster_ip (ocf:heartbeat:IPaddr2): Started node-02
    * web-service (systemd:nginx): Started node-02
  * Clone Set: dummy-check-clone [dummy-check]:
    * Started: [ node-01 node-03 ]
  * Clone Set: stateful-demo-clone [stateful-demo] (promotable):
    * Promoted: [ node-03 ]
    * Unpromoted: [ node-01 ]

Older releases may display Masters and Slaves instead of Promoted and Unpromoted.