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

[HTCondor-users] trigger STARTD_CRON_NODEHEALTH_EXECUTABLE at condor start



Hi,

we do use the STARTD_CRON feature excessively and there is one problem, if we set the Period to short the load on the bigger machines gets too high due to the startd publishing all the news over all the slots (consumption_policy = true) and also the fs-checks are not meant to run in a minut intervall. 

Hence when the condor service starts as the the STARTD_CRON seems to wait for the period instead of running once right away some variables that get set by the STARTD_CRON are unset which leads to problems. 

On the other hand as the period in our case is 900sec it seems not to be desirable to set positive defaults on all probes for this time and wait for the first run of STARTD_CRON. 

It would be very handy me think if the STARTD_CRON_NODEHEALTH_EXECUTABLE would run once right after the start-up of condor, maybe there is a way to trigger this even ? 

If that is more complicated than I think it should be I guess persistent classadds are aother way to go for these probes ? 

Best
christoph

-- 
Christoph Beyer
DESY Hamburg
IT-Department

Notkestr. 85
Building 02b, Room 009
22607 Hamburg

phone:+49-(0)40-8998-2317
mail: christoph.beyer@xxxxxxx