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

Re: [Condor-users] DAGMan



Hi Matt,

> Ok, I give the =!= a try tomorrow, when I'm back at work. But 
> in my opinion,
> subnet is defined automatically by Condor on every node.

I replaced "!=" with "=!=" in the APPEND_REQUIREMENTS-expression and now
everything works fine, thank you so much!
But, I don't understand why this helped me out. I'm still convinced that
there must be some inconsistency with the dagman and APPEND_REQUIREMENTS. As
I said, when I submit 'normal' jobs, I don't need the "=!=" in the
APPEND_REQUIREMENTS-expression, just for the dagman.
Correct me if I'm wrong, but as far as I understand, "A=!=B" means that this
expression is only compared, if A is defined. I swear, subnet is defined on
every machine, that's a standard condor setting. (By the way, I tried
APPEND_REQUIREMENTS=Mips>1000 to check, if there is a problem when
APPEND_REQUIREMENTS is too long, but this didn't work either.)

Do you have any idea?

Kind regards,
Thomas Bauer

  
> > >> So you know, condor_dagman actually condor_submits jobs. So 
> > > 
> > > Yeah, I know that ;-) And at this point I don't understand 
> > why the dag-jobs
> > > don't start. When I submit 'normal' jobs from this node, 
> > everything works
> > > fine, but dag-jobs produce the error-message, that the 
> > requirements couldn't
> > > be evaluated.
> > > 
> > >> if you are 
> > >> submitting the DAG from host X to host Y, condor_dagman will be 
> > >> submitting on host Y with host Y's configuration.
> > > 
> > > I would say this is partially correct, because the 
> > scheduler makes the
> > > problems here, named by you as host X. The schedlog on host 
> > X tells about
> > > the problems to evaluate the requirements. And when I change the
> > > configuration on host X to "APPEND_REQUIREMENTS =" (in 
> > order to shrink the
> > > requirements) the dag-jobs run as expected.
> > > 
> > > Any ideas?
> > 
> > What version of Condor? I can take your requirements 
> expression above 
> 
> I must admit 'only' Condor 7.0.1 on the submit and execute 
> node. I've read
> the release notes for the recent versions, but there were no 
> comments about
> fixing a problem in windows-dagman which is similar to my problem.
> 
> > and run a job with it (after chaning the OpSys to "Linux").
> 
> I don't know, maybe this is only a problem in the 
> Windows-Version of DAGman?
> Something like "640K ought to be enough for anybody"? ;-)
> 
> Kind regards,
> Thomas Bauer
>  
> > e.g.
> > 
> > universe = vanilla
> > executable = /bin/true
> > Requirements = (subnet != "10.4.130" && subnet != "10.4.131" 
> > && subnet 
> > != "10.4.133") && (Arch == "INTEL") && (OpSys == "Linux") 
> && (Disk >= 
> > DiskUsage) && ((Memory * 1024) >= ImageSize)
> > queue 1
> > 
> > 
> > Best,
> > 
> > 
> > matt
> > 
> > > Kind regards, 
> > > Thomas Bauer
> > > 
> > > 
> > > 
> > >> Thomas Bauer wrote:
> > >>> Hello,
> > >>>
> > >>> I searched myself for the error and I think I found it. I 
> > >> would say there is
> > >>> a bug in the dag-system.
> > >>>
> > >>> Our machines are configured in the condor_config with the entry
> > >>>   APPEND_REQUIREMENTS =
> > >>> subnet!="10.4.130"&&subnet!="10.4.131"&&subnet!="10.4.133" 
> > >>> because I don't want jobs to be started in these subnets.
> > >>>
> > >>> As I posted earlier, when I submit the dag-job with this 
> > >> configuration, I
> > >>> get the error 
> > >>>   The Requirements attribute for job 10.0 did not evaluate. 
> > >> Unable to start
> > >>> job
> > >>>
> > >>> Now, when I add
> > >>>   APPEND_REQUIREMENTS =  
> > >>> to the condor_config.local on the submitting node 
> > >> everything works fine. Is
> > >>> there a rigorouser limitation for the requirements in the 
> > >> dag-system? 
> > >>> BTW, there is no probleam with the longer 
> > >> APPEND_REQUIREMENTS-setting when I
> > >>> submit normal jobs.
> > >>>
> > >>> Kind regards,
> > >>> Thomas Bauer
> > >>>
> > >>>
> > >>>
> > >>>
> > >>>> -----Ursprüngliche Nachricht-----
> > >>>> Von: condor-users-bounces@xxxxxxxxxxx 
> > >>>> [mailto:condor-users-bounces@xxxxxxxxxxx] Im Auftrag von 
> > >> Thomas Bauer
> > >>>> Gesendet: Dienstag, 14. Oktober 2008 12:19
> > >>>> An: 'Condor-Users Mail List'
> > >>>> Betreff: [Condor-users] DAGMan
> > >>>>
> > >>>> Hi folks,
> > >>>>
> > >>>> I try to submit a DAG-Job from a Windows-System to a 
> > >> Windows-System.
> > >>>> Unfortunately the DAG-jobs don't start. It says in the 
> SchedLog:
> > >>>>
> > >>>> 10/14 12:07:45 The Requirements attribute for job 10.0 did 
> > >>>> not evaluate.
> > >>>> Unable to start job
> > >>>>
> > >>>> Does anybody know where I have to look for the problem? The 
> > >>>> accordingly
> > >>>> SUB-Files have the entry
> > >>>>
> > >>>> Requirements = (OpSys=="WINNT51"&&Arch=="INTEL")
> > >>>>
> > >>>> Kind regards,
> > >>>> Thomas Bauer
> > >>>>
> > >>>> _______________________________________________
> > >>>> Condor-users mailing list
> > >>>> To unsubscribe, send a message to 
> > >>>> condor-users-request@xxxxxxxxxxx with a
> > >>>> subject: Unsubscribe
> > >>>> You can also unsubscribe by visiting
> > >>>> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> > >>>>
> > >>>> The archives can be found at: 
> > >>>> https://lists.cs.wisc.edu/archive/condor-users/
> > >>>>
> > >>> _______________________________________________
> > >>> Condor-users mailing list
> > >>> To unsubscribe, send a message to 
> > >> condor-users-request@xxxxxxxxxxx with a
> > >>> subject: Unsubscribe
> > >>> You can also unsubscribe by visiting
> > >>> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> > >>>
> > >>> The archives can be found at: 
> > >>> https://lists.cs.wisc.edu/archive/condor-users/
> > >> _______________________________________________
> > >> Condor-users mailing list
> > >> To unsubscribe, send a message to 
> > >> condor-users-request@xxxxxxxxxxx with a
> > >> subject: Unsubscribe
> > >> You can also unsubscribe by visiting
> > >> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> > >>
> > >> The archives can be found at: 
> > >> https://lists.cs.wisc.edu/archive/condor-users/
> > >>
> > > 
> > > _______________________________________________
> > > Condor-users mailing list
> > > To unsubscribe, send a message to 
> > condor-users-request@xxxxxxxxxxx with a
> > > subject: Unsubscribe
> > > You can also unsubscribe by visiting
> > > https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> > > 
> > > The archives can be found at: 
> > > https://lists.cs.wisc.edu/archive/condor-users/
> > _______________________________________________
> > Condor-users mailing list
> > To unsubscribe, send a message to 
> > condor-users-request@xxxxxxxxxxx with a
> > subject: Unsubscribe
> > You can also unsubscribe by visiting
> > https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> > 
> > The archives can be found at: 
> > https://lists.cs.wisc.edu/archive/condor-users/
> > 
> 
> _______________________________________________
> Condor-users mailing list
> To unsubscribe, send a message to 
> condor-users-request@xxxxxxxxxxx with a
> subject: Unsubscribe
> You can also unsubscribe by visiting
> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> 
> The archives can be found at: 
> https://lists.cs.wisc.edu/archive/condor-users/
>