Bug 1628659

vg_access_mode is a required input parameter.  lvm_validate() just checks that the VGs are consistent with the vg_access_mode.

lvm_validate() also switches the var VG_access_mode from the string "lvmlockd" to the required int 1, so that it calls the proper ${access_mode}_deactivate() in stop. Otherwise, if it remains "lvmlockd", instead of either 1,2,3,4, you'll end up with the "is not properly configured in cluster. It's unsafe!"

# lvm_validate:
lvm_validate() {
        [...]
        case ${VG_access_mode} in
        lvmlockd)
                VG_access_mode=1



More info on how the resources were set up, w/ vg_access_mode having being provided:

pcs resource create lvm1 --group HA_LVM1 ocf:heartbeat:LVM-activate lvname="ha" vgname="HARDING1" activation_mode=exclusive vg_access_mode=lvmlockd
pcs resource create ha1 --group HA_LVM1 Filesystem device="/dev/HARDING1/ha" directory="/mnt/ha1" fstype="xfs" "options=noatime" op monitor interval=10s  
pcs constraint order start lvmlockd-clone then HA_LVM1

pcs resource create lvm2 --group HA_LVM2 ocf:heartbeat:LVM-activate lvname="ha" vgname="HARDING2" activation_mode=exclusive vg_access_mode=lvmlockd
pcs resource create ha2 --group HA_LVM2 Filesystem device="/dev/HARDING2/ha" directory="/mnt/ha2" fstype="xfs" "options=noatime" op monitor interval=10s  
pcs constraint order start lvmlockd-clone then HA_LVM2



[root@harding-02 ~]# pcs config
Cluster Name: HARDING
Corosync Nodes:
 harding-02 harding-03
Pacemaker Nodes:
 harding-02 harding-03

Resources:
 Clone: dlm_for_lvmlockd-clone
  Meta Attrs: interleave=true ordered=true 
  Resource: dlm_for_lvmlockd (class=ocf provider=pacemaker type=controld)
   Operations: monitor interval=30s (dlm_for_lvmlockd-monitor-interval-30s)
               start interval=0s timeout=90 (dlm_for_lvmlockd-start-interval-0s)
               stop interval=0s timeout=100 (dlm_for_lvmlockd-stop-interval-0s)
 Clone: lvmlockd-clone
  Meta Attrs: interleave=true ordered=true 
  Resource: lvmlockd (class=ocf provider=heartbeat type=lvmlockd)
   Attributes: with_cmirrord=1
   Operations: monitor interval=30s (lvmlockd-monitor-interval-30s)
               start interval=0s timeout=90s (lvmlockd-start-interval-0s)
               stop interval=0s timeout=90s (lvmlockd-stop-interval-0s)
 Group: HA_LVM1
  Meta Attrs: target-role=Stopped 
  Resource: lvm1 (class=ocf provider=heartbeat type=LVM-activate)
   Attributes: activation_mode=exclusive lvname=ha vg_access_mode=lvmlockd vgname=HARDING1
   Operations: monitor interval=30s timeout=90s (lvm1-monitor-interval-30s)
               start interval=0s timeout=90s (lvm1-start-interval-0s)
               stop interval=0s timeout=90s (lvm1-stop-interval-0s)
  Resource: ha1 (class=ocf provider=heartbeat type=Filesystem)
   Attributes: device=/dev/HARDING1/ha directory=/mnt/ha1 fstype=xfs options=noatime
   Operations: monitor interval=10s (ha1-monitor-interval-10s)
               notify interval=0s timeout=60s (ha1-notify-interval-0s)
               start interval=0s timeout=60s (ha1-start-interval-0s)
               stop interval=0s timeout=60s (ha1-stop-interval-0s)
 Group: HA_LVM2
  Meta Attrs: target-role=Stopped 
  Resource: lvm2 (class=ocf provider=heartbeat type=LVM-activate)
   Attributes: activation_mode=exclusive lvname=ha vg_access_mode=lvmlockd vgname=HARDING2
   Operations: monitor interval=30s timeout=90s (lvm2-monitor-interval-30s)
               start interval=0s timeout=90s (lvm2-start-interval-0s)
               stop interval=0s timeout=90s (lvm2-stop-interval-0s)
  Resource: ha2 (class=ocf provider=heartbeat type=Filesystem)
   Attributes: device=/dev/HARDING2/ha directory=/mnt/ha2 fstype=xfs options=noatime
   Operations: monitor interval=10s (ha2-monitor-interval-10s)
               notify interval=0s timeout=60s (ha2-notify-interval-0s)
               start interval=0s timeout=60s (ha2-start-interval-0s)
               stop interval=0s timeout=60s (ha2-stop-interval-0s)

Stonith Devices:
 Resource: smoke-apc (class=stonith type=fence_apc)
  Attributes: delay=5 ipaddr=smoke-apc login=apc passwd=apc pcmk_host_check=static-list pcmk_host_list=harding-02,harding-03 pcmk_host_map=harding-02:3;harding-03:4 switch=1
  Operations: monitor interval=60s (smoke-apc-monitor-interval-60s)
Fencing Levels:

Location Constraints:
Ordering Constraints:
  start lvmlockd-clone then start HA_LVM1 (kind:Mandatory)
  start lvmlockd-clone then start HA_LVM2 (kind:Mandatory)
  Resource Sets:
    set dlm_for_lvmlockd-clone lvmlockd-clone sequential=true
Colocation Constraints:
Ticket Constraints:

Alerts:
 No alerts defined

Resources Defaults:
 No defaults set
Operations Defaults:
 No defaults set

Cluster Properties:
 cluster-infrastructure: corosync
 cluster-name: HARDING
 dc-version: 1.1.19-7.el7-c3c624ea3d
 have-watchdog: false
 last-lrm-refresh: 1536851845
 no-quorum-policy: freeze

Quorum:
  Options:

It's hard to understand how this agent wasn't sanity checked when it was originally written.  Also, the code for handling the access mode is rather sloppy, overwriting the string value with a numeric value in the same variable.  The processing of input parameters should be done outside of lvm_validate since stop/status shouldn't be doing the rest of the validate function.

Created attachment 1483124 [details]
UNTESTED diff

This diff illustrates the kind of changes I think make sense here.  It's untested so will need some verification.

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2018:3278