[HTCondor-users] Problem with executors

[Joaquin Sanroman <joaquin.sanroman@xxxxxxxxxxx>]

Hi,

I have a problem with mi condor executors. The main problem is that when I shutdown all daemons in one executor machine, the executor instances still appears when execute condor_status from any other machine.

I have tried to shutdown all startd daemons using condor_off -all -startd and all instances dissapears except the instances from the machine which first shutdown. For you to understand this is the behaviour:

This is our pool:

Now i will kill my daemons on "macfer" machine:

The next step is to check if instances have dissapeared with condor_status command:

As you can see, they are still there. Now we will try to disconect the other slots with condor_off -all -startd:

Obviously, canÂt connect to master macfer.local because all daemons are died. Now, if we do a condor_status, the only machine which is connected (not really connected) is macfer.local:

Â

After this, this machine doesnÂt dissapear althought it is not enable to execute jobs.

Does anybody know how to fix this problem? because i wouldnÂt like this instances appears in the pool

Our condor version are 8.0.6 and 8.0.7 (macfer is the only machine with 8.0.7)

Kind Regards.