How do I check if there is an already running dask scheduler?

Question

I want to start a local cluster from python with a specific number of workers, and then connect a client to it.

cluster = LocalCluster(n_workers=8, ip='127.0.0.1')
client = Client(cluster)

But before, I want to check if there is an existing local cluster, started for example by the dask-scheduler command. Is there a way to do this ?

score 7 · Accepted Answer · answered Apr 13 '18 at 15:57

7

There is no standard convention to check if a scheduler exists on your machine. The best you can do is try with a short timeout. The default port is 8786

from dask.distributed import Client, TimeoutError

try:
    client = Client('tcp://localhost:8786', timeout='2s')
except TimeoutError:
    pass

answered Apr 13 '18 at 15:57

MRocklin

55,641
23
163
235

Thanks, this was exactly what I am looking for! – slaw Sep 21 '18 at 17:43

How do I check if there is an already running dask scheduler?

1 Answers1