[DRMAA-WG] Minutes from Conference call - Jun 23th

[Daniel Templeton]

On 7/12/10 1:04 PM, Peter Tröger wrote:
>
> Am 07.07.2010 um 20:52 schrieb Daniel Templeton:
>
>> And I still don't see any resolution for the problem of getting back a
>> list of several hundred thousand jobs.  I think we've had this
>> discussion before.  Anyone have notes on what we decided?
>
> We also discussed that, and basically agreed on trusting in the OO 
> language's scalability capabilities.

That's what I was afraid of.  Let me go stew about that a little, and 
I'll be ready to argue about it on the next call. ;)

Daniel

>
> Best,
> Peter.
>
>
>>
>> Daniel
>>
>> On 07/06/10 15:16, Peter Tröger wrote:
>>> Hi,
>>>
>>>>> 2. Monitoring jobs not submitted by DRMAA - final decision
>>>>
>>>> - Important to be able to obtain information about non-DRMAA
>>>>  jobs which have not ended.
>>>>
>>>> - Desirable to obtain information about non-DRMAA jobs which
>>>>  have recently ended, where "recently" is DRM-defined.
>>>>
>>>> - Importance (and ability to obtain) information about jobs
>>>>  submitted by others unclear.  Agreed that if there is support
>>>>  for obtaining information about other's jobs, must support a
>>>>  filter to limit to current user's jobs.
>>>>
>>>> - Desirable to obtain information about non-DRMAA jobs using
>>>>  APIs similar to those used for jobInfo.  Use Dictionary type
>>>>  to support DRM-specific info.
>>>
>>> After some thinking, I decided to put that into the JobSession
>>> interface:
>>>
>>> === snip ===
>>> interface JobSession {
>>>   readonly attribute string contact;
>>>   readonly attribute string sessionName;
>>>   Job getJob(string jobId);
>>>   sequence<Job>  getSessionJobs(JobInfo filter);
>>>   sequence<Job>  getAllJobs(JobInfo filter);
>>>   Job runJob(in DRMAA::JobTemplate jobTemplate) raises (???);
>>> ...
>>> === snip ===
>>> Non-session jobs are identified by not having the Job::session
>>> attribute set. I am still unsure about the "filtering for current
>>> user" part, since this would bring us the concept of  user name, which
>>> was omitted so far.
>>>
>>> Best,
>>> Peter.
>>>
>>>
>>>>
>>>>> 5. Cleaning up the spreadsheet:
>>>>
>>>> No discussion.
>>>> -- 
>>>>  drmaa-wg mailing list
>>>>  drmaa-wg at ogf.org
>>>>  http://www.ogf.org/mailman/listinfo/drmaa-wg
>>>
>>> -- 
>>>   drmaa-wg mailing list
>>>   drmaa-wg at ogf.org
>>>   http://www.ogf.org/mailman/listinfo/drmaa-wg
>> -- 
>>  drmaa-wg mailing list
>>  drmaa-wg at ogf.org
>>  http://www.ogf.org/mailman/listinfo/drmaa-wg
>