[saga-rg] SAGA conference call
Graeme Pound
G.E.POUND at soton.ac.uk
Thu Apr 20 08:21:13 CDT 2006
Andre,
The client must be able to specify the resource manager to which jobs
are submitted, as a service URL or machine name. This should be a
argument for either the JobService constructor, or the
JobService.createJob() & JobService.runJob() methods.
I guess that conceptually I prefer specifying the resource manager
during the construction of the JobService object. However in an earlier
mail Chris Smith indicated that the 'host' argument of the runJob()
method was originally intended to indicate the machine upon which the
job should be run (mapping to an element of the SAGA_HostList
attribute). This is confusing and the 'host' argument should be removed,
or redefined to indicate the resource manager.
Graeme
Andre Merzky wrote:
> No, unfortunetly not. I added an issue in the issue list
> (forgot to do that before). I think that, if we don't find
> an agreement by GGF17 (in two weeks), we just decide on some
> approach, e.g. always refereing to the resource manager,
> until we know where that breaks.
>
> What would you preferred solution be? Sorry if I should
> know that from earlier mails, I left that issue resting as
> well *sigh*
>
> Cheers, Andre.
>
>
>
> Quoting [Graeme Pound] (Apr 20 2006):
>> Date: Thu, 20 Apr 2006 13:41:35 +0100
>> From: Graeme Pound <G.E.POUND at soton.ac.uk>
>> To: Andre Merzky <andre at merzky.net>
>> CC: Shantenu Jha <s.jha at ucl.ac.uk>, saga-rg at ggf.org
>> Subject: Re: [saga-rg] SAGA conference call
>>
>> Andre,
>>
>> Has there been any resolution about the issues raised in the thread
>> "JobService.submitJob() query"
>> (http://www-unix.gridforum.org/mail_archive/saga-rg/2006/03/maillist.html).
>>
>> I am a little out of the loop on this.
>>
>> Thanks,
>>
>> Graeme
>>
>>
>>
>>
>> Andre Merzky wrote:
>>> Hi folx,
>>>
>>> attached is the current version of the issue list.
>>>
>>> Cheers, Andre.
>>>
>>>
>>> Quoting [Andre Merzky] (Apr 19 2006):
>>>> Date: Wed, 19 Apr 2006 21:24:19 +0200
>>>> From: Andre Merzky <andre at merzky.net>
>>>> To: Shantenu Jha <s.jha at ucl.ac.uk>
>>>> Cc: saga-rg at ggf.org
>>>> Subject: Re: [saga-rg] SAGA conference call
>>>>
>>>> Hi folx,
>>>>
>>>> here are the call details:
>>>>
>>>> Date: 20 April 2006
>>>> 2:00 pm London
>>>> Time: 3:00 pm Amsterdam
>>>> 8:00 am Baton Rouge
>>>> 10:00 pm Tokyo
>>>> Number: 0844.888.8888 (UK)
>>>> 01805.004.102 (Germany)
>>>> +44.870.088.5706 (Others)
>>>>
>>>> Code: 808044
>>>>
>>>> Cheers, Andre.
>>>>
>>>>
>>>> Quoting [Shantenu Jha] (Apr 18 2006):
>>>>> Date: Tue, 18 Apr 2006 14:14:58 +0100 (BST)
>>>>> From: Shantenu Jha <s.jha at ucl.ac.uk>
>>>>> To: saga-rg at ggf.org
>>>>> Subject: [saga-rg] SAGA conference call
>>>>>
>>>>> Hi All,
>>>>>
>>>>> After a brief hiatus, we're ready to resume our regular conference
>>>>> calls.
>>>>>
>>>>> Suggested time for the next call is 2PM BST on Thursday 20th April, which
>>>>> translates to 8AM CST and 3PM in Amsterdam (and probably close to
>>>>> midnight
>>>>> in Tokyo. Sorry Pascal.)
>>>>>
>>>>> The agenda for the next call is:
>>>>> - Issue List
>>>>> - SAGA and GridRPC
>>>>> How to coordinate activities of GridRPC into SAGA.
>>>>> - SAGA and GridCPR
>>>>> - AOB ?
>>>>>
>>>>> Conference call details to come soon.
>>>>>
>>>>> If you'd like to join, but suggested time doesn't work for you, please
>>>>> let us know.
>>>>>
>>>>> Thanks.
>>>>>
>>>>> ------------------------------------------------------------------------
>>>>>
>>>>>
>>>>> +-------------------------------------------------------------+
>>>>>
>>>>> #####
>>>>> # # ##### ###### # #
>>>>> # # # # # ## #
>>>>> # # # # ##### # # #
>>>>> # # ##### # # # #
>>>>> # # # # # ##
>>>>> ##### # ###### # #
>>>>> ###
>>>>> # #### #### # # ###### ####
>>>>> # # # # # # #
>>>>> # #### #### # # ##### ####
>>>>> # # # # # # #
>>>>> # # # # # # # # # #
>>>>> ### #### #### #### ###### ####
>>>>>
>>>>> +-------------------------------------------------------------+
>>>>>
>>>>>
>>>>> 12) change signature of read/write to have len_out as out param, and
>>>>> to return
>>>>> error code (== EAGAIN on unsucessful read)
>>>>> - DECIDED at GGF16
>>>>> - is that _really_ the right way?
>>>>> - OPEN, URGENT
>>>>>
>>>>> 29) check if bulk operations can be done with task containers
>>>>> - explicit support needed/wanted for argument list generation?
>>>>> - consider parametrisable job descriptions for bulk jobs?
>>>>> - TODO STEPHAN
>>>>>
>>>>> 30) ACLs!
>>>>> - Later, after we get input from the security area and GFS
>>>>> - we actually got that input for files/name spaces, so that
>>>>> should be done!
>>>>> - OPEN, URGENT, MAJOR
>>>>>
>>>>> 36) - examples are not normative for language binding
>>>>> - provide one examples in various languages
>>>>> - TODO TOM: Fortran
>>>>> - TODO HARTMUT: Python
>>>>>
>>>>> 47) check JSDL compliance
>>>>> - TODO CHRIS/ANDRE, URGENT
>>>>>
>>>>> 55) check strawman for references
>>>>> - OPEN
>>>>>
>>>>> 60) add attributes to other objects? consistency model might be an
>>>>> option for replication, blocking/nonblocking for file I/O etc.
>>>>> - OPEN
>>>>>
>>>>> 70) describe task model verbosely in the spec, leave details to
>>>>> language bindings.
>>>>> As we have now job implementing the task interface, that should
>>>>> be reconsidered!
>>>>> - OPEN
>>>>>
>>>>> 76) add specs for literals (attribute and metric names),
>>>>> and add them like enums
>>>>> - OPEN, URGENT
>>>>>
>>>>> 77) add details for getURL and getName
>>>>> - TODO ANDRE, URGENT
>>>>>
>>>>> 81) document pattern and extended read/write
>>>>> - TODO ANDRE, URGENT
>>>>>
>>>>> 82) Explain sidl.SIDLException !
>>>>> - OPEN, URGENT
>>>>>
>>>>> 83) Exceptions should go into the SIDL part!
>>>>> - would clutter the SIDL?
>>>>> - OPEN
>>>>>
>>>>> 88) for job: add clean_up (can also be called after termination)
>>>>> - OPEN, URGENT
>>>>>
>>>>> 90) exceptions missing in job!
>>>>> - OPEN, URGENT
>>>>>
>>>>> 93) introduction needs sanity checks in respect to the groups
>>>>> development
>>>>> (WG/RG split)
>>>>> - OPEN
>>>>>
>>>>> 95) Security on observables missing: need callout on metric
>>>>> read/write
>>>>> - sync with stream auth example
>>>>> - OPEN, URGENT
>>>>>
>>>>> 97) flags (and their inheritance) seems somewhat confusing
>>>>> - see name space, reuse values in file etc
>>>>> - TODO ANDRE
>>>>>
>>>>> 98) LogicalFiles/Directories: define attributes
>>>>> - TODO ANDRE
>>>>>
>>>>> 102) document what classes need to implement ErrorHandler
>>>>> (shouldn't that be all?)
>>>>> - OPEN
>>>>>
>>>>> 108) explain why we don't specify constness. Or should we?
>>>>> - OPEN, URGENT
>>>>>
>>>>> 109) Formatting should be more TeX like
>>>>> - TODO ANDRE
>>>>>
>>>>> 111) describe that file constructors etc. perform open
>>>>> - TODO ANDRE
>>>>>
>>>>> 112) monitoring needs detailed descriptions.
>>>>> - TODO ANDRE, URGENT
>>>>>
>>>>> 115) detailed description for name_space methods out of sync
>>>>> ns_entry methods are missing.
>>>>> - TODO ANDRE, URGENT
>>>>>
>>>>> 118) stream: detailed description out of sync
>>>>> - TODO ANDRE, URGENT
>>>>>
>>>>> 119) task: detailed description out of sync
>>>>> - TODO ANDRE, URGENT
>>>>>
>>>>> 125) monitoring as in streams ok, use in all objects, define
>>>>> available metrics
>>>>> - TODO ANDRE, URGENT
>>>>>
>>>>> 126) complete descriptions of available metrics
>>>>> - see 125
>>>>> - TODO ANDRE, URGENT
>>>>>
>>>>> 130) clearify the error handler interface for async calls
>>>>> - OPEN, URGENT
>>>>>
>>>>> 131) clearify what a NULL string/object is we sometimes return
>>>>> - OPEN, URGENT
>>
>
>
>
More information about the saga-rg
mailing list