Re: [HTCondor-users] question about accounting groups

[Stefano Dal Pra <stefano.dalpra@xxxxxxxxxxxx>]

Hi,
On 29/04/21 14:39, Jeff Templon wrote:
> Hi,
>
> Thanks Stefano.
>
> So what I am calling âaccountingâ, HTCondor is calling âhistoryâ, that 
> helps find things in the manual.
>
> One thing you say comes back to an earlier bit of the discussion:
>
> On 29 Apr 2021, at 14:16, Stefano Dal Pra wrote:
>
> > for this to work:
> > on each schedd you have defined PER_JOB_HISTORY_DIR to an existing 
> > directory
> > in that directory you'll find one file per finished job (such as 
> > history.2411281.0)
> >
> > Alternatively, on each schedd you run condor_history, almost the same 
> > way:
>
> Both options start with âon each scheddâ. So the Central Manager does 
> not keep track of history, it only keeps track of aggregate usage in 
> the sense explained in the User Priority parts of the HTCondor 
> manual. Correct?
Yes, correct. for (every following wrong assumption is of my own only) 
scalability and robustness reasons the CM is not aware of several "gory 
details",
like the complete set of job classads for every job in the whole pool 
(or cluster, as I used to say back in the LSF days).
The schedd babysits every job submitted to him through its whole life, 
knows about its evolution from one state to the next
(idle, running, hold|done|removed) and is the one who can be considered 
"accountable" for those; in that sense it looks ok that it is the one 
who knows about
job accounting (which is, btw, the complete set of job classads when it 
reached its final status).

Stefano


> When I logged into our Central Manager and ran âcondor_historyâ there 
> was no result.
>
> JT