FENCE_SCSI RESOURCE CREATION: unknown error

Hi All,

I am trying to create a resource of fence_scsi resource agent but each time I am getting error:

First i tried:

sudo crm configure primitive FENCE_SCSI stonith:fence_scsi params action=off devices="/dev/sdb" verbose=false op monitor interval=60s timeout=0s

got error: Sep 13 05:07:17 [30109] ESDNAS1 stonith-ng: info: update_remaining_timeout: Attempted to execute agent fence_scsi (monitor) the maximum number of times (2) allowed
Sep 13 05:07:17 [30109] ESDNAS1 stonith-ng: notice: log_operation: Operation ‘monitor’ [19853] for device ‘FENCE_SCSI’ returned: -201 (Generic Pacemaker error)
Sep 13 05:07:17 [30109] ESDNAS1 stonith-ng: warning: log_operation: FENCE_SCSI:19853 [ Failed: nodename or key is required ]
Sep 13 05:07:17 [30109] ESDNAS1 stonith-ng: warning: log_operation: FENCE_SCSI:19853 [ ]
Sep 13 05:07:17 [30109] ESDNAS1 stonith-ng: warning: log_operation: FENCE_SCSI:19853 [ Please use ‘-h’ for usage ]
Sep 13 05:07:17 [30109] ESDNAS1 stonith-ng: warning: log_operation: FENCE_SCSI:19853 [ ]
Sep 13 05:07:17 [30110] ESDNAS1 lrmd: notice: log_finished: finished - rsc:FENCE_SCSI action:start call_id:86 exit-code:1 exec-time:1180ms queue-time:0ms

Secondly I tried giving nodename as an option:

sudo crm configure primitive FENCE_SCSI stonith:fence_scsi params action=off devices="/dev/sdb" verbose=false nodename=linux-3wc8 op monitor interval=60s timeout=0s

Still the resource is not getting created.

Failed Actions:

  • FENCE_SCSI_start_0 on linux-3wc8 ‘unknown error’ (1): call=78, status=Error, exitreason=‘none’,
    last-rc-change=‘Mon Sep 30 03:07:09 2019’, queued=0ms, exec=1202ms
  • FENCE_SCSI_start_0 on linux-fqft ‘unknown error’ (1): call=85, status=Error, exitreason=‘none’,
    last-rc-change=‘Mon Sep 30 03:07:13 2019’, queued=1ms, exec=1199ms

The overall config is as follows:

# crm config show
node 183725435: linux-3wc8
node 183725437: linux-fqft
node 183725442: linux-tzkc
node 183725451: linux-7rbk
primitive FENCE_SCSI stonith:fence_scsi \
params action=off devices="/dev/sdb" verbose=false nodename=linux-3wc8 \
op monitor interval=60s timeout=0s
primitive PGSQL_FS Filesystem \
params device="/dev/sdb" directory="/nas" fstype=ext3 \
op monitor interval=30 timeout=60
primitive SERVICE systemd:sample \
op start timeout=120 interval=0 \
op stop timeout=120 interval=0 \
op monitor timeout=100s interval=10s
primitive failover-ip IPaddr \
params ip=172.18.64.0 nic=eth0 cidr_netmask=16 \
op monitor interval=10s
property cib-bootstrap-options: \
have-watchdog=false \
dc-version=1.1.13-10.4-6f22ad7 \
cluster-infrastructure=corosync \
cluster-name=hacluster \
stonith-enabled=false \
placement-strategy=balanced
rsc_defaults rsc-options: \
resource-stickiness=1 \
migration-threshold=3
op_defaults op-options: \
timeout=600 \
record-pending=true

Can someone please help with this!
I am stuck at this step.

  1. Never use naming like /dev/sdb - this is not a persistent name and might change in future reboots - you can use /dev/disk/by-id/scsi-…
    If your device is under multipath -use multipath device without friendly names (unless you have mapped wwids to friendly name and the setup is consistent on both nodes).
    For multiple devices you can use “,” to separate them.
  2. Is your storage allowing you to use SCSI persistent reservations?
  3. As the error states you lack some vital information for the fencing mechanism to work
    Check all options via:
crm ra info stonith:fence_scsi

The data bellow is based on memory and could be wrong , check all options via the ‘crm ra’ command above!

So you might need pcmk_host_list=“node1 node2 nodeN”
All nodes should be listed here.
pcmk_reboot_action=“off”
This allows us to ‘off’ the node instead of a reboot.
meta provides=“unfencing”
This tells stonith-ng to “on” action on unfence.