Suggestions for verifying changes.

-------------------------------------------------------------------

sanlock: use helper to set max_sectors_kb

1. Do lockstart on a sanlock-based shared VG

[root@rhel9a ~]# vgchange --lockstart bb
  Skipping global lock: lockspace not found or started
  VG bb starting sanlock lockspace
  Starting locking.  Waiting for sanlock may take 20 sec to 3 min...

2. Look up the major:minor number of the internal lvmlock LV

[root@rhel9a ~]# ls -lL /dev/mapper/bb-lvmlock
brw-rw---- 1 root disk 253, 2 Dec 12 11:54 /dev/mapper/bb-lvmlock

3. Check that max_sectors_kb is set to 1024 for the lvmlock device

[root@rhel9a ~]# cat /sys/dev/block/253:2/queue/max_sectors_kb
1024

-------------------------------------------------------------------

watchdog timeout configuration 

1. Verify that the watchdog device on the system supports a 30 second timeout

[root@rhel9a ~]# wdmd -t 30 /dev/watchdog
/dev/watchdog0 gettimeout 30
/dev/watchdog0 settimeout 30 result 30
/dev/watchdog0 gettimeout 30

2. Edit /etc/sanlock/sanlock.conf to set:

watchdog_fire_timeout = 30

3. Do lockstart on a sanlock-based shared VG

[root@rhel9a ~]# vgchange --lockstart bb
  Skipping global lock: lockspace not found or started
  VG bb starting sanlock lockspace
  Starting locking.  Waiting for sanlock may take 20 sec to 3 min...

4. Verify that wdmd and sanlock daemons show they are using '30'

[root@rhel9a ~]# wdmd -d | grep 'fire_timeout 30'
wdmd 4655 socket_gid 179 high_priority 0 now 88663 last_keepalive 88663 last_closeunclean 0 allow_scripts 0 kill_script_sec 0 fire_timeout 30

[root@rhel9a ~]# sanlock status -D | grep 'fire_timeout=30'
    watchdog_fire_timeout=30

--------------------------------------------------------------------

wdmd: close watchdog when not used

1. Start and stop a sanlock-based shared VG

[root@rhel9a ~]# vgchange --lockstart bb
  Skipping global lock: lockspace not found or started
  VG bb starting sanlock lockspace
  Starting locking.  Waiting for sanlock may take 20 sec to 3 min...
[root@rhel9a ~]# vgchange --lockstop bb

2. Check that the watchdog device is not open

[root@rhel9a ~]# lsof /dev/watchdog0

3. Start the VG

[root@rhel9a ~]# vgchange --lockstart bb
  Skipping global lock: lockspace not found or started
  VG bb starting sanlock lockspace
  Starting locking.  Waiting for sanlock may take 20 sec to 3 min...

4. Check that the watchdog device is open

[root@rhel9a ~]# lsof /dev/watchdog0
COMMAND  PID USER   FD   TYPE DEVICE SIZE/OFF NODE NAME
wdmd    4655 root    9w   CHR  247,0      0t0  715 /dev/watchdog0

5. Stop the VG

[root@rhel9a ~]# vgchange --lockstop bb

6. Check that the watchdog device is not open

[root@rhel9a ~]# lsof /dev/watchdog0

--------------------------------------------------------------------

sanlock: skip short delay in delta lease reacquire
sanlock: use product_uuid for host name

1. Start and stop a sanlock-based shared VG

[root@rhel9a ~]# vgchange --lockstart bb
  Skipping global lock: lockspace not found or started
  VG bb starting sanlock lockspace
  Starting locking.  Waiting for sanlock may take 20 sec to 3 min...
[root@rhel9a ~]# vgchange --lockstop bb

2. Record the sanlock host name

(compare with this later)

[root@rhel9a ~]# sanlock status | grep daemon
daemon a95d7bae-cbcd-4d08-bcae-51d21e54cdda.rhel9a

2. Restart the lvmlockd and sanlock daemons

[root@rhel9a ~]# killall lvmlockd
[root@rhel9a ~]# sanlock shutdown -f 1
shutdown force 1 wait 0
shutdown done 0
[root@rhel9a ~]# sanlock daemon
[root@rhel9a ~]# lvmlockd

3. Check the sanlock host name

(should be the same as above)

[root@rhel9a ~]# sanlock status | grep daemon
daemon a95d7bae-cbcd-4d08-bcae-51d21e54cdda.rhel9a

4. Time starting the VG

The result should be about 4 seconds, not 20 seconds,
because the same host name was reused.

[root@rhel9a ~]# time vgchange --lockstart bb
  Skipping global lock: lockspace not found or started
  VG bb starting sanlock lockspace
  Starting locking.  Waiting for sanlock may take 20 sec to 3 min...

real    0m4.051s
user    0m0.006s
sys     0m0.013s