[ClusterLabs] unable to start fence_scsi on a new add node

[Andrei Borzenkov]

16.04.2020 18:58, Stefan Sabolowitsch пишет:
> Hi there,
> i have expanded a cluster with 2 nodes with an additional one "elastic-03". However, fence_scsi does not start on the new node.
> 
> pcs-status:
> [root at logger cluster]# pcs status
> Cluster name: cluster_elastic
> Stack: corosync
> Current DC: elastic-02 (version 1.1.20-5.el7_7.2-3c4c782f70) - partition with quorum
> Last updated: Thu Apr 16 17:38:16 2020
> Last change: Thu Apr 16 17:23:43 2020 by root via cibadmin on elastic-03
> 
> 3 nodes configured
> 10 resources configured
> 
> Online: [ elastic-01 elastic-02 elastic-03 ]
> 
> Full list of resources:
> 
>  scsi	(stonith:fence_scsi):	Stopped
>  Clone Set: dlm-clone [dlm]
>      Started: [ elastic-01 elastic-02 ]
>      Stopped: [ elastic-03 ]
>  Clone Set: clvmd-clone [clvmd]
>      Started: [ elastic-01 elastic-02 ]
>      Stopped: [ elastic-03 ]
>  Clone Set: fs_gfs2-clone [fs_gfs2]
>      Started: [ elastic-01 elastic-02 ]
>      Stopped: [ elastic-03 ]
> 
> Failed Fencing Actions:
> * unfencing of elastic-03 failed: delegate=, client=crmd.5149, origin=elastic-02,
>     last-failed='Thu Apr 16 17:23:43 2020'
> 
> Daemon Status:
>   corosync: active/enabled
>   pacemaker: active/enabled
>   pcsd: active/enabled
> 
> 
> corosync.log 
> Apr 16 17:27:10 [4572] logger stonith-ng:   notice: can_fence_host_with_device:       scsi can fence (off) elastic-01 : static-list
> Apr 16 17:27:12 [4572] logger stonith-ng:   notice: can_fence_host_with_device:       scsi can fence (off) elastic-02 : static-list
> Apr 16 17:27:13 [4572] logger stonith-ng:   notice: can_fence_host_with_device:       scsi can not fence (off) elasti c-03: static-list
> Apr 16 17:38:43 [4572] logger stonith-ng:   notice: can_fence_host_with_device:       scsi can not fence (on) elastic -03: static-list

You probably need to update your stonith resource to include new node.

> Apr 16 17:38:43 [4572] logger stonith-ng:   notice: remote_op_done:   Operation on of elastic-03 by <no-one> for crmd .5149 at elastic-02.4b624305: No such device
> Apr 16 17:38:43 [4576] logger.feltengroup.local       crmd:    error: tengine_stonith_notify:   Unfencing of elastic-03 by <anyone> failed: No such device (-19)
> 
> [root at logger cluster]# stonith_admin -L
>  scsi
> 1 devices found
> 
> [root at logger cluster]# stonith_admin -l elastic-03
> No devices found
> 
> Thanks for any help here.
> Stefan
> 
> _______________________________________________
> Manage your subscription:
> https://lists.clusterlabs.org/mailman/listinfo/users
> 
> ClusterLabs home: https://www.clusterlabs.org/
>