Memory Leak observed when using sbd watchdog type stonith

Issue

• While using sbd watchdog type fencing, the sbd: watcher: Pacemaker thread grows steadily over time, and is never decreased until cluster stop or system reboot:

  Mon Jun 17 12:01:01 PM EEST 2024 VVV
  USER         PID    PPID %CPU %MEM    VSZ   RSS TT       STAT  STARTED     TIME WCHAN                            COMMAND
  root         914       1  0.0  0.2  20808 20456 ?        SL     Jun 14 00:01:13 do_sigtimedwait                  sbd: inquisitor
  root         915     914  0.2  5.5 450088 449472 ?       SL     Jun 14 00:07:50 do_poll.constprop.0              sbd: watcher: Pacemaker <<<===
  root         916     914  0.0  0.5  43288 42884 ?        SL     Jun 14 00:00:50 do_poll.constprop.0              sbd: watcher: Cluster
  
  Mon Jun 17 12:01:21 PM EEST 2024
  USER         PID    PPID %CPU %MEM    VSZ   RSS TT       STAT  STARTED     TIME WCHAN                            COMMAND
  root         914       1  0.0  0.2  20808 20456 ?        SL     Jun 14 00:01:13 do_sigtimedwait                  sbd: inquisitor
  root         915     914  0.2  5.5 450236 449728 ?       SL     Jun 14 00:07:50 do_poll.constprop.0              sbd: watcher: Pacemaker  <<<===
  root         916     914  0.0  0.5  43288 42884 ?        SL     Jun 14 00:00:50 do_poll.constprop.0              sbd: watcher: Cluster
  

• The rate of growth is very slow, and usually observed at about 128 kb every ~95s ( or ~120 Mb ) a day. If issue persists too long though, this can lead to low memory conditions and memory pressure, and even out of memory ( "OOM" ) events.