[ClusterLabs] Pacemaker / ubuntu doesn't see my sbd device: what am I missing?

[Ken Gaillot]

With watchdog-only SBD you don't need a fence agent; it's built-in to
Pacemaker when you set the stonith-watchdog-timeout cluster property.

However watchdog-only SBD isn't sufficient for a 2-node cluster,
because each node will assume the other self-fences but neither will.
You need either a shared disk or true quorum (via a third node or
corosync-qdevice).

On Wed, 2022-04-06 at 19:34 +0000, Tavanyar, Simon wrote:
> This is my first time using Pacemaker, and I wanted to try watchdog-
> only fencing with SBD.
> I’m running on Ubuntu 21.10 and Pacemaker v2.0.5
> My cluster is up just fine with Dummy services on two nodes.
> Systemd says my sbd device is active and running.
> But the ‘stonith’ command that Pacemaker uses won’t find it, so the
> resource fails to start in the cluster.
>  
> Help much appreciated!
> Thanks
> Simon
>  
>  
>  
> $ sudo stonith -t external/sbd -E -S
> external/sbd[361914]: ERROR: No sbd device(s) found in the
> configuration.
> WARN: external_status: 'sbd status' failed with rc 1
> ERROR: external/sbd device not accessible.
>  
>  
> $ systemctl status sbd
> ● sbd.service - Shared-storage based fencing daemon
> Loaded: loaded (/lib/systemd/system/sbd.service; enabled; vendor
> preset: enabled)
> Active: active (running) since Fri 2022-04-01 15:18:04 EDT; 4 days
> ago
> Docs: man:sbd(8)
> Process: 2474278 ExecStart=/usr/sbin/sbd $SBD_OPTS -p
> /var/run/sbd.pid watch (code=exited, status=0/SUCCESS)
> Main PID: 2474279 (sbd)
> Tasks: 3 (limit: 38258)
> Memory: 11.2M
> CPU: 4min 7.329s
> CGroup: /system.slice/sbd.service
> ├─2474279 sbd: inquisitor
> ├─2474280 sbd: watcher: Pacemaker
> └─2474281 sbd: watcher: Cluster
>  
>  
> $ sudo pcs status
> Cluster name: Axx
> Cluster Summary:
>   * Stack: corosync
>   * Current DC: node0 (version 2.0.5-ba59be7122) - partition with
> quorum
>   * Last updated: Wed Apr  6 14:38:44 2022
>   * Last change:  Wed Apr  6 14:38:35 2022 by root via cibadmin on
> node0
>   * 2 nodes configured
>   * 6 resource instances configured
>  
> Node List:
>   * Online: [ node0 node1 ]
>  
> Full List of Resources:
>   * Resource Group: AxxDummy:
>     * p_Dummy_1 (ocf::heartbeat:Dummy):  Started node0
>     * p_Dummy_2 (ocf::heartbeat:Dummy):  Started node0
>     * p_Dummy_3 (ocf::heartbeat:Dummy):  Started node0
>     * ClusterIP (ocf::heartbeat:IPaddr2):        Started node0
>   * p_Dummy_4   (ocf::heartbeat:Dummy):  Started node0
>   * fence-sbd   (stonith:external/sbd):  Stopped
>  
> Failed Resource Actions:
>   * fence-sbd_start_0 on node0 'error' (1): call=51,
> status='complete', exitreason='', last-rc-change='2022-04-06 14:38:13
> -04:00', queued=0ms, exec=3102ms
>   * fence-sbd_start_0 on node1 'error' (1): call=41,
> status='complete', exitreason='', last-rc-change='2022-04-06 14:38:09
> -04:00', queued=0ms, exec=3094ms
>  
> Daemon Status:
>   corosync: active/enabled
>   pacemaker: active/enabled
>   pcsd: active/enabled
>   sbd: active/enabled
>  
>  
> This is from /var/log/syslog
> Apr 6 14:40:43 ubuntuserver pacemaker-controld[349716]: notice:
> Requesting local execution of start operation for fence-sbd on node0
> Apr 6 14:40:43 ubuntuserver external/sbd[349924]: [349930]: ERROR: No
> sbd device(s) found in the configuration.
> Apr 6 14:40:46 ubuntuserver pacemaker-fenced[349712]: notice:
> Operation 'monitor' [349931] for device 'fence-sbd' returned: -61 (No
> data available)
> Apr 6 14:40:46 ubuntuserver pacemaker-fenced[349712]: warning: fence-
> sbd:349931 [ Performing: stonith -t external/sbd -E -S ]
> Apr 6 14:40:46 ubuntuserver pacemaker-fenced[349712]: warning: fence-
> sbd:349931 [ failed: 1 ]
>  
> _______________________________________________
> Manage your subscription:
> https://lists.clusterlabs.org/mailman/listinfo/users
> 
> ClusterLabs home: https://www.clusterlabs.org/
-- 
Ken Gaillot <kgaillot at redhat.com>