Client hangs if cluster is closed/shutdown

[mccullerlp]

I find that there is no way to query if the cluster is still live, and any tasks submitted will hang with no method for checking connection status.

Is there any workaround or heartbeat that occurs to prevent long hangs?

[slishak]

I'm having this problem too. Can easily reproduce:

cmd:

ipcluster start -n=4

Python terminal:

import ipyparallel as ipp
c = ipp.Client

cmd:

ipcluster stop

Python terminal:

c.queue_status()

causes an infinite hang.