Re: [HTCondor-users] condor_defrag does not start to defrag

[John M Knoeller <johnkn@xxxxxxxxxxx>]

One suggestion on this -

The Defrag daemon will pick a random STARTD to drain if all of the candidates have the same DEFRAG_RANK
so it might be wise to have a RANK expression that puts the STARTDs in to buckets rather than one that 
is continuously variable.

I don't have anything better to suggest at this time. 

-tj

-----Original Message-----
From: Carsten Aulbert <carsten.aulbert@xxxxxxxxxx> 
Sent: Wednesday, December 2, 2020 10:32 AM
To: HTCondor-Users Mail List <htcondor-users@xxxxxxxxxxx>; John M Knoeller <johnkn@xxxxxxxxxxx>
Subject: Re: [HTCondor-users] condor_defrag does not start to defrag

Additional finding - probably not that surprising especially due to low 
interval values:

With the default setting of

DEFRAG_RANK = -ExpectedMachineGracefulDrainingBadput

the same machine(s) seems to be drained again and again. Thus, we will 
try to factor in "LastDrainStartTime" as well to fix this and report 
back - unless someone already has a good setting here and would like to 
share it :)

Cheers

Carsten