Hello,

I found something in the status output of sinfo to be peculiar.

$ sinfo
PARTITION AVAIL  TIMELIMIT  NODES  STATE NODELIST
ubique*      up   infinite      1  down* cal-059
ubique*      up   infinite      5  drain cal-[172-176]
ubique*      up   infinite    132    mix 
cal-[005-027,029-032,042-054,056-058,060-074,079-086,089-128,143,145-146,148-149,151-171]
ubique*      up   infinite     37  alloc 
cal-[001-004,033-041,055,075-078,087-088,129-142,144,147,150]
ubique*      up   infinite      1   down cal-028

Here we can see cal-059, cal-[172-176] and cal-028 are either in state down or 
drain.

So, let's list only the nodes which are down,
$ sinfo -t down
PARTITION AVAIL  TIMELIMIT  NODES  STATE NODELIST
ubique*      up   infinite      1  down* cal-059
ubique*      up   infinite      4  drain cal-[173-176]
ubique*      up   infinite      1   down cal-028

All of a sudden, one node disappeared from the list - cal-172. Have I stumbled 
upon a bug?

However, it is shown again if I ask for only the "drain" state,
$ sinfo -t drain
PARTITION AVAIL  TIMELIMIT  NODES  STATE NODELIST
ubique*      up   infinite      5  drain cal-[172-176]

$ sinfo --version
slurm 14.11.4

I am unsure if this happens in 14.11.5, but we haven't upgraded to that just 
yet.

If there's anything I can do for verification, please don't hesitate to get in 
touch.

Wbr
Andreas
________________________________


Confidentiality Notice: This message is private and may contain confidential 
and proprietary information. If you have received this message in error, please 
notify us and remove it from your system and note that you must not copy, 
distribute or take any action in reliance on it. Any unauthorized use or 
disclosure of the contents of this message is not permitted and may be unlawful.

Reply via email to