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

Re: [HTCondor-users] Condor is still writing temp data to Condor is still writing temp data to /var/opt/condor even though I specified otherwise



Thanks. They were not overwritten, Ive already triple checked that. Can I run “condor_config_val -v execute” from my head node? That is the only one I am able to log into (I don’t know the IP addresses of the compute nodes).


Nate Mobley

Millennium Engineering & Integration Company

ISSO/Systems Administrator

Desk: (256) 489-7847

Cell (Voice Only): (256) 655-5570

nmobley@xxxxxxxxxxxxxx

www.meicompany.com


On Jan 8, 2018, at 16:22, Todd Tannenbaum <tannenba@xxxxxxxxxxx> wrote:

On 1/8/2018 4:15 PM, Mobley, Nate (Millennium) wrote:
Thanks for the input. Yes, I've restarted the system (head node plus 9 compute nodes) as well as the Condor service, multiple times. I originally made this change in May, and it seemed to help for a while.

Maybe the config files were overwritten?

The EXECUTE config setting is used on your compute nodes.

If you login to one of your compute nodes and enter the following command, HTCondor will tell you that value of the EXECUTE config knob and also the file/line number where it is defined:

  condor_config_val -v execute

Knowing the file where HTCondor is reading EXECUTE from could be helpful.

See "man condor_config_val" for more about this command-line tool to query about configuration values.

regards,
Todd



Nate Mobley
Millennium Engineering & Integration Company
ISSO/Systems Administrator
Desk: (256) 489-7847
Cell (Voice Only): (256) 655-5570
MEI Help Desk:  (703) 413-7771
nmobley@xxxxxxxxxxxxxx
www.meicompany.com
-----Original Message-----
From: Todd Tannenbaum [mailto:tannenba@xxxxxxxxxxx]
Sent: Monday, January 08, 2018 4:12 PM
To: HTCondor-Users Mail List <htcondor-users@xxxxxxxxxxx>; Mobley, Nate (Millennium) <nmobley@xxxxxxxxxxxxxx>
Subject: Re: [HTCondor-users] Condor is still writing temp data to Condor is still writing temp data to /var/opt/condor even though I specified otherwise
On 1/8/2018 3:30 PM, Mobley, Nate (Millennium) wrote:
"/var/opt/condor/execute" is on a very small partition with less than
1 gb available (This is how the system was configured when I hired on
to this position, and I cannot change this partition size at this
time). My customer's runs are crashing when Condor tries to write any
significant amount of data to this location, according to the log
file. I changed the EXECUTE path to "/export/condor/execute" in both
config files ("/opt/condor/etc/condor_config" and
"/opt/condor/etc/condor_config.local"), but for some reason Condor is
still trying to write data to "/var/opt/condor/execute."
After changing the EXECUTE path in the condor_config file(s), did you
restart the condor service?  (i.e. as root do /sbin/service condor stop,
then /sbin/service condor start?)
The HTCondor daemons do not continuously poll the config files for
changes; most times when you make a change in the config file, you need
to tell HTCondor to re-read it either via "condor_reconfig" if it is a
setting that can be changed on-the-fly, or via restarting the condor
service (killing all condor daemons and restarting them) if the setting
cannot be changed on the fly.  Most settings that involve filesystem
paths cannot be changed on-the-fly.
regards and hope this helps,
Todd


--
Todd Tannenbaum <tannenba@xxxxxxxxxxx> University of Wisconsin-Madison
Center for High Throughput Computing   Department of Computer Sciences
HTCondor Technical Lead                1210 W. Dayton St. Rm #4257
Phone: (608) 263-7132                  Madison, WI 53706-1685