Node returns drained and can be resumed (or cancelled) by NHC

mej · May 22, 2020 · 3df3e9c · 3df3e9c
1 parent 7486112
commit 3df3e9c
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 3 deletions.
diff --git a/helpers/node-mark-online b/helpers/node-mark-online
@@ -60,7 +60,7 @@ elif [[ "$NHC_RM" == "slurm" ]]; then
  # SLURM does not run the HealthCheckProgram on nodes in the DOWN state,
  # but if someone runs NHC by hand, we want to be able to do the right thing.
  case "$STATUS" in
- down*|drain*|drng*|fail*|maint*)
+ down*|drain*|drng*|fail*|maint*|boot*)
  # If there is no old note, and we've not been told to ignore that, do not online the node.
  if [[ "$OLD_NOTE_LEADER" == "none" && "$IGNORE_EMPTY_NOTE" != "1" ]]; then
  echo "$0: Not onlining $HOSTNAME: No note set."

diff --git a/helpers/node-mark-reboot b/helpers/node-mark-reboot
@@ -25,7 +25,7 @@ NOTE="$*"
 if [[ "$NHC_RM" == "slurm" ]]; then
  SLURM_SINFO="${SLURM_SINFO:-sinfo}"
  SLURM_SCONTROL="${SLURM_SCONTROL:-scontrol}"
- SLURM_SC_REBOOT_ARGS="${SLURM_SC_REBOOT_ARGS:-reboot ASAP NextState=RESUME}"
+ SLURM_SC_REBOOT_ARGS="${SLURM_SC_REBOOT_ARGS:-reboot ASAP NextState=DOWN}"
 
  LINE=( $($SLURM_SINFO -o '%t %E' -hn $HOSTNAME) )
  STATUS="${LINE[0]}"
@@ -47,7 +47,6 @@ if [[ "$NHC_RM" == "slurm" ]]; then
  if [[ "$OLD_NOTE_LEADER" != "none" && "$OLD_NOTE_LEADER" != "$LEADER" ]]; then
  LEADER="$OLD_NOTE_LEADER"
  NOTE="$OLD_NOTE"
- SLURM_SC_REBOOT_ARGS="reboot ASAP NextState=DOWN"
  fi
  echo "$0: Marking $STATUS $HOSTNAME for reboot: $LEADER $NOTE"
  exec $SLURM_SCONTROL $SLURM_SC_REBOOT_ARGS Reason="$LEADER $NOTE" $HOSTNAME