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

Re: [HTCondor-users] efficient way to connect to collector and scheduler

Absent the Python bindings, given your languages of choice, you might consider running a condor_status or _q to pull the raw classads for everything youâre interested in, and use condor_status and condor_q against that file to gather the specifics. I have a queue watchdog script that takes this approach â a single query gets a big file with everything I need, then the condor_q formatting commands I use to generate the e-mails donât have to bother the schedd.


In the meantime, make sure people are not using condor_q | grep. When you specify constraints and filters, the schedd can sort that out in memory, rather than sending a bunch of useless stuff that will be ignored by the grep.


Michael V Pelletier

Principal Engineer

Raytheon Technologies

Information Technology

Digital Transormation & Innovation


From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> On Behalf Of rmorgan466@xxxxxxxxx
Sent: Friday, September 25, 2020 7:58 AM
To: HTCondor-Users Mail List <htcondor-users@xxxxxxxxxxx>
Subject: [External] [HTCondor-users] efficient way to connect to collector and scheduler


i use condor_status and condor_q to get metrics. In the newer version of HTCondor is there a way to directly connect to a socket/port to pull metrics. We have 30+ users who constantly use condor_q and condor_status. I would like to write a microservice to get all this data most efficiently. Language of choice here is: Java and Go.



--- Get your facts first, then you can distort them as you please.--