Re: [HTCondor-users] more information on incorrect RemoteUserCpu and -currentrun

[Jeff Templon <templon@xxxxxxxxx>]

Hi Cole,

> On 16 Dec 2022, at 23:46, Cole Bollig via HTCondor-users <htcondor-users@xxxxxxxxxxx> wrote:
> 
> Hi Jeff,
> 
> In the eyes of Condor, the current run time of a job is how long that it has been in the running state for.Which the job enters once resources have been claimed and a shadow starts. So, current time is more specifically of how long the shadow has been running as opposed to "slot connect duration". This why you saw a run time similar to when in the job log a reconnect occurred since the log message before hand stated the shadow died.
> 

Thanks for the answer.  This should be clarified in the HTCondor documentation, where it says

RUN_TIME
(Non-batch mode only) Wall-clock time accumulated by the job to date in days, hours, minutes, and seconds.

Your answer says that what HTCondor calls âRUN_TIMEâ is how long the current instance of the jobâs shadow has been running.  This is not the same as the time accumulated by âthe jobâ from a user point of view, the user being in this point the target audience of the condor_q documentation.  I only rarely care how long a shadow has been running; I almost always want to know how long the work I submitted has been running.

JT