[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HTCondor-users] schedd on every submit node VS single schedd



Iâve found that having a single central schedd works better for me - everythingâs visible in one place, users donât get as confused while logged in to multiple machines, etc. Thereâs some constraints in that a shared filesystem across all submitters and execute nodes is needed for best results, but thatâs the way everything was originally set up before HTCondor anyway.

 

There used to be scalability considerations - one group of my users was submitting tens of thousands of jobs at a time and before the batched output for condor_q came out, and before the CERN-driven scalability tuning a couple years ago, things would get a bit unruly when they got busy and I wound up having to give them their own schedd, but thatâs no longer a big issue.

 

Michael V. Pelletier
Information Technology
Digital Transformation & Innovation
Integrated Defense Systems
Raytheon Company

 

From: HTCondor-users [mailto:htcondor-users-bounces@xxxxxxxxxxx] On Behalf Of David Cohen
Sent: Monday, November 19, 2018 6:36 AM
To: condor-users@xxxxxxxxxxx
Subject: [External] [HTCondor-users] schedd on every submit node VS single schedd

 

Hi,

I'm starting an HTCondor system for the first time, to replace a torque/maui cluster, and trying to figure out the basic topology before I start.

What is the benefit of having a schedd on every submit node vs single schedd.

 

Is there a good reason not to have just:

1 htcondor-ce host

1 central manager host

1 schedd host

n startd hosts

 

David