[Bigjob-users] question

[Andre Luckow]

Hi Pradeep,
the only workaround to this 30 minutes termination issue is to start
the BJ manager on another machine. You can do so by copying the Grid
proxy certificate to another machine. The only pre-requisite is that
the Globus client tools and SAGA are installed on this machine.

I hope we can provide a disconnected mode of operation in the near
future (maybe in the context of TROY).

Best,
Andre

On Sat, Oct 22, 2011 at 4:13 PM, pradeep kumar Mantha
<pradeepm66 at gmail.com> wrote:
> Hi!
>  I am working with Ryky to use bigjob for his simulations. I think, many
> people mentioned this problem in the Radical meeting. Just to rephrase his
> question, He is using  bigjob example with globus resource url on oliver
> machine. He is using head node just to run his example( python
> example_local_generic.py). On the other nodes, running examples is not
> possible, since initiating grid-proxy is failing. If it is a simulation
> executes > 30 mins then it causes a problem since the process ( python
> example_local_generic.py ) running on the head node. It might get
> terminated. Is there any alternative for this?
>
> - thanks
> pradeep
>
> On Sat, Oct 22, 2011 at 4:06 AM, Andre Luckow <aluckow at cct.lsu.edu> wrote:
>>
>> Hi Ryky,
>> in order to distribute jobs to a cluster you have several options. You
>> can either utilize the SAGA/Globus, SAGA/PBS or SAGA/SSH:
>>
>> Following the example linked on the BigJob page:
>>
>> https://svn.cct.lsu.edu/repos/saga-projects/applications/bigjob/trunk/generic/examples/example_local_generic.py
>>
>> you can use the following lrms_urls:
>>
>> lrms_url = "gram://oliver1.loni.org/jobmanager-pbs" # globus resource
>> url used when globus is used. (LONI)
>> lrms_url = "pbspro://oliver1.loni.org" # pbspro resource url used when
>> pbspro scheduling system is used.(Futuregrid or LSU Machines)
>> lrms_url = "ssh://oliver2.loni.org"
>>
>> Of course, the respective SAGA adaptors need to be installed.
>>
>> Best,
>> Andre
>>
>>
>> On Thu, Oct 20, 2011 at 11:38 PM, Ryky Nelson <rnels12 at tigers.lsu.edu>
>> wrote:
>> > Hi,
>> > I was successfully executing big-job example in oliver1 which is a
>> > headnode.
>> > However, because it's a headnode I wanted to try executing the job on
>> > oliver2 but it didn't work. Is there any alternative for that?
>> > Thank you.
>> > Best,
>> > Ryky
>> >
>> > _______________________________________________
>> > Bigjob-users mailing list
>> > Bigjob-users at mail.cct.lsu.edu
>> > https://mail.cct.lsu.edu/mailman/listinfo/bigjob-users
>> >
>> >
>> _______________________________________________
>> Bigjob-users mailing list
>> Bigjob-users at mail.cct.lsu.edu
>> https://mail.cct.lsu.edu/mailman/listinfo/bigjob-users
>
>