Hi Steve,
I've disabled the proxy purger on one of my CEs we'll see if that one still
creates jobs that ends up in 'W' start or starts crashing.
If the worst comes to the worst I can always write a cron job to delete
proxies that expired more than N days ago.
Yours,
Chris.
> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
> [log in to unmask]] On Behalf Of Stephen Jones
> Sent: 01 February 2012 12:00
> To: [log in to unmask]
> Subject: Re: Ops meeting at 11am
>
> Hi Chris,
>
> I've never tried it. Which of us should be the guinea pig?!
>
> For what it is worth, I assume for a minute that the setting entirely
> disables the proxy purger. That means no proxy would ever be purged,
> which (at least) solves the specific problem (stagein file missing)
> while generating others (no purging whatsoever, proxy files stick
> around) and jobs starting with a stale proxy file in situ.
>
> Having said that, I think that behaviour is better than the current
> behavior which results in W jobs.
>
> Is there any real problem with stale proxies remaining in situ? It
> seems untidy, and a waste of space, but so what? If so, we could do
> with a delegation_purge_rate setting (which controls how often the
> purger runs) _and_ a delegation_purge_delay setting (which controls
> how stale a proxy has to be before it gets the chop).
> Set delegation_purge_rate to "now and again" and set
> delegation_purge_delay to two weeks, and that's that, as far as I can
> see.
>
> Steve
>
> Chris Brew wrote:
> > Hi,
> >
> > Reading through the ticket, I see this:
> >
> >
> http://grid.pd.infn.it/cream/field.php?n=Main.HowToConfigureTheProxyPu
> > rger
> >
> > where it says that setting delegation_purge_rate=-1 in
> > cream-config.xml disables the proxy purger. Has anyone tried that?
> > Does it fix this problem/cause other problems?
> >
> > Yours,
> > Chris.
> >
> >
> >> -----Original Message-----
> >> From: Testbed Support for GridPP member institutes [mailto:TB-
> >> [log in to unmask]] On Behalf Of Stephen Jones
> >> Sent: 01 February 2012 10:07
> >> To: [log in to unmask]
> >> Subject: Re: Ops meeting at 11am
> >>
> >> Yes, https://ggus.eu/tech/ticket_show.php?ticket=72506
> >>
> >> They'll first fix the bugs which are either easy or which generate
> >> the most complaints.
> >>
> >> Could you corroborate the bug by adding your observations?
> >>
> >> Cheers,
> >>
> >> Steve
> >>
> >>
> >>
> >>
> >>
> >> Chris Brew wrote:
> >>
> >>> Is there a bug open with the CreamCE developers about that? They
> >>> should be cleanly aborted/deleted from the batch system but we
> >>>
> >> usually
> >>
> >>> have hundreds of the damn things clogging up our batch system.
> >>>
> >>> Chris.
> >>>
> >>>
> >>>
> >>>> -----Original Message-----
> >>>> From: Testbed Support for GridPP member institutes [mailto:TB-
> >>>> [log in to unmask]] On Behalf Of Stephen Jones
> >>>> Sent: 31 January 2012 17:00
> >>>> To: [log in to unmask]
> >>>> Subject: Re: Ops meeting at 11am
> >>>>
> >>>> Hi,
> >>>>
> >>>> The main problem that biomed encounters at our site is laid out in
> >>>> this
> >>>> bug:
> >>>>
> >>>> https://ggus.eu/tech/ticket_show.php?ticket=72506
> >>>>
> >>>> Some jobs get stuck in W state because they arrived with a proxy
> >>>>
> >> that
> >>
> >>>> was too short for the delay that occurred before they got to run.
> >>>>
> >>>> There's not much I can do about that.
> >>>>
> >>>> Steve
> >>>>
> >>>>
> >>>> Stephen Burke wrote:
> >>>>
> >>>>
> >>>>> Testbed Support for GridPP member institutes [mailto:TB-
> >>>>>
> >>>>>
> >>>>>
> >>>>>> [log in to unmask]] On Behalf Of Daniela Bauer said:
> >>>>>> They have a nagios and file tickets more or less automatically.
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>> They are also learning how things work, I've had several
> >>>>>
> >> discussions
> >>
> >>>> with Franck Michel about the info system and he started asking
> >>>> questions *before* he started submitting tickets!
> >>>>
> >>>>
> >>>>> Stephen
> >>>>>
> >>>>>
> >>>>>
> >>>> --
> >>>> Steve Jones [log in to unmask]
> >>>> System Administrator office: 220
> >>>> High Energy Physics Division tel (int): 42334
> >>>> Oliver Lodge Laboratory tel (ext): +44 (0)151 794
> >>>>
> >> 2334
> >>
> >>>> University of Liverpool
> >>>> http://www.liv.ac.uk/physics/hep/
> >>>>
> >>>>
> >> --
> >> Steve Jones [log in to unmask]
> >> System Administrator office: 220
> >> High Energy Physics Division tel (int): 42334
> >> Oliver Lodge Laboratory tel (ext): +44 (0)151 794
> 2334
> >> University of Liverpool
> >> http://www.liv.ac.uk/physics/hep/
> >>
>
>
> --
> Steve Jones [log in to unmask]
> System Administrator office: 220
> High Energy Physics Division tel (int): 42334
> Oliver Lodge Laboratory tel (ext): +44 (0)151 794 2334
> University of Liverpool
> http://www.liv.ac.uk/physics/hep/
|