airavata-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From K Yoshimoto <kenn...@sdsc.edu>
Subject [marpierc@iu.edu: Re: [SciGaP-Dev] CIPRES-Airavata integration proposal]
Date Thu, 03 Jul 2014 17:48:32 GMT

Posting this also to dev@airavata, at Marlon's request.

On 7/3/14, 1:39 PM, K Yoshimoto wrote:
> Another question.  I was able to create a project and experiment,
> then launch the experiment.  It was EXECUTING, but is now SUSPENDED.
> How should the gateway developer investigate the cause of suspension?
> I think that job state would not normally happen on the Trestles resource,
> at least.
> 
> php createExperiment.php
> php createExperiment.php <username> <experiment_name> <project_ID>
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php createExperiment.php kenneth kennethexp1 kennethtest_a26036de-d58c-4f6c-b38d-e60d30c13704
> object(Airavata\Model\Workspace\Experiment\Experiment)#9 (18) {
>    ["experimentID"]=>
>    string(21) "DO_NOT_SET_AT_CLIENTS"
>    ["projectID"]=>
>    string(48) "kennethtest_a26036de-d58c-4f6c-b38d-e60d30c13704"
>    ["creationTime"]=>
>    NULL
>    ["userName"]=>
>    string(7) "kenneth"
>    ["name"]=>
>    string(11) "kennethexp1"
>    ["description"]=>
>    NULL
>    ["applicationId"]=>
>    string(11) "SimpleEcho2"
>    ["applicationVersion"]=>
>    NULL
>    ["workflowTemplateId"]=>
>    NULL
>    ["workflowTemplateVersion"]=>
>    NULL
>    ["userConfigurationData"]=>
>    object(Airavata\Model\Workspace\Experiment\UserConfigurationData)#6 (7) {
>      ["airavataAutoSchedule"]=>
>      int(0)
>      ["overrideManualScheduledParams"]=>
>      int(0)
>      ["shareExperimentPublicly"]=>
>      bool(false)
>      ["computationalResourceScheduling"]=>
>      object(Airavata\Model\Workspace\Experiment\ComputationalResourceScheduling)#4 (10)
{
>        ["resourceHostId"]=>
>        string(17) "trestles.sdsc.edu"
>        ["totalCPUCount"]=>
>        int(1)
>        ["nodeCount"]=>
>        int(1)
>        ["numberOfThreads"]=>
>        int(0)
>        ["queueName"]=>
>        string(6) "normal"
>        ["wallTimeLimit"]=>
>        int(15)
>        ["jobStartTime"]=>
>        int(0)
>        ["totalPhysicalMemory"]=>
>        int(0)
>        ["computationalProjectAccount"]=>
>        NULL
>        ["ComputationalProjectAccount"]=>
>        string(6) "sds128"
>      }
>      ["advanceInputDataHandling"]=>
>      NULL
>      ["advanceOutputDataHandling"]=>
>      NULL
>      ["qosParams"]=>
>      NULL
>    }
>    ["workflowExecutionInstanceId"]=>
>    NULL
>    ["experimentInputs"]=>
>    array(1) {
>      [0]=>
>      object(Airavata\Model\Workspace\Experiment\DataObjectType)#7 (4) {
>        ["key"]=>
>        string(10) "echo_input"
>        ["value"]=>
>        string(23) "echo_output=Hello World"
>        ["type"]=>
>        int(0)
>        ["metaData"]=>
>        NULL
>      }
>    }
>    ["experimentOutputs"]=>
>    array(1) {
>      [0]=>
>      object(Airavata\Model\Workspace\Experiment\DataObjectType)#8 (4) {
>        ["key"]=>
>        string(11) "echo_output"
>        ["value"]=>
>        string(0) ""
>        ["type"]=>
>        int(0)
>        ["metaData"]=>
>        NULL
>      }
>    }
>    ["experimentStatus"]=>
>    NULL
>    ["stateChangeList"]=>
>    NULL
>    ["workflowNodeDetailsList"]=>
>    NULL
>    ["errors"]=>
>    NULL
> }
> Experiment kennethexp1_7364a59d-4bd0-4cd8-9f59-82a39863df92 created!
> 
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php launchExperiment.php
> Please provide an experimentID.
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php launchExperiment.php kennethexp1_7364a59d-4bd0-4cd8-9f59-82a39863df92
> Experiment kennethexp1_7364a59d-4bd0-4cd8-9f59-82a39863df92 is launched.
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php getExperimentStatus.php kennethexp1_7364a59d-4bd0-4cd8-9f59-82a39863df92
> experiment status = EXECUTING
> <br>
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php getExperimentStatus.php kennethexp1_7364a59d-4bd0-4cd8-9f59-82a39863df92
> experiment status = SUSPENDED
> <br>
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php getExperimentStatus.php kennethexp1_7364a59d-4bd0-4cd8-9f59-82a39863df92
> experiment status = SUSPENDED
> <br>
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php getExperimentStatus.php kennethexp1_7364a59d-4bd0-4cd8-9f59-82a39863df92
> experiment status = SUSPENDED
> <br>
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php getExperimentStatus.php kennethexp1_7364a59d-4bd0-4cd8-9f59-82a39863df92
> experiment status = SUSPENDED
> <br>
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php getExperimentStatus.php kennethexp1_7364a59d-4bd0-4cd8-9f59-82a39863df92
> experiment status = SUSPENDED
> <br>
> john@buffy:/media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
> 
> On Thu, Jul 03, 2014 at 10:23:14AM -0700, K Yoshimoto wrote:
>> That works now.  Thanks!
>> 
>> On Thu, Jul 03, 2014 at 05:00:21PM +0000, Singh, Raminder jeet wrote:
>>> Hi Kenneth,
>>> 
>>> With help from a Student in Sri Lanka we are able to identify and fix the issue.
Issue was related to the socket send timeout. I updated the clients to add following property.
Please take a git pull and test.
>>> 
>>> $transport->setSendTimeout($airavataconfig['AIRAVATA_TIMEOUT']);
>>> 
>>> Thanks
>>> Raminder
>>> 
>>> On Jul 3, 2014, at 12:17 AM, K Yoshimoto <kenneth@sdsc.edu> wrote:
>>> 
>>>> Still timed out.
>>>> 
>>>> On Thu, Jul 03, 2014 at 04:06:42AM +0000, Singh, Raminder jeet wrote:
>>>>> Hi Kenneth
>>>>> 
>>>>> Try to increase the AIRAVATA_TIMEOUT = 5000 to 20000 (in ms) in airavata-client-properties.ini.
Its the connection timeout connecting to the server.
>>>>> 
>>>>> Thanks
>>>>> Raminder
>>>>> 
>>>>> On Jul 2, 2014, at 11:57 PM, K Yoshimoto <kenneth@sdsc.edu> wrote:
>>>>> 
>>>>>> I tried the new clients (had to set servername from localhost to
>>>>>> the iu server).  I'm still seeing a timeout:
>>>>>> 
>>>>>> airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples$
php getAPIVersion.php
>>>>>> PHP Fatal error:  Uncaught exception 'Thrift\Exception\TException'
with message 'TSocket: Could not connect to gw111.iu.xsede.org:8930 (Connection timed out
[110])' in /media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/lib/Thrift/Transport/TSocket.php:235
>>>>>> Stack trace:
>>>>>> #0 /media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples/getAPIVersion.php(39):
Thrift\Transport\TSocket->open()
>>>>>> #1 {main}
>>>>>> thrown in /media/mdhpss/scigap/phpclientsamples/airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/lib/Thrift/Transport/TSocket.php
on line 235
>>>>>> 
>>>>>> 
>>>>>> On Wed, Jul 02, 2014 at 09:40:39PM -0400, Marlon Pierce wrote:
>>>>>>> Hi Kenneth--
>>>>>>> 
>>>>>>> I did a quick successful check with the Airavata php clients
(in airavata/airavata-api/airavata-client-sdks/airavata-php-sdk/src/main/resources/php-cli-samples)
>>>>>>> (see other email for details).
>>>>>>> 
>>>>>>> Please use these instead of the obsolete PHP client samples.
>>>>>>> 
>>>>>>> Thanks--
>>>>>>> 
>>>>>>> Marlon
>>>>>>> 
>>>>>>> On 7/2/14, 7:04 PM, K Yoshimoto wrote:
>>>>>>>> I've started playing with the php client samples.
>>>>>>>> I'm trying to create project, but running into some kind
>>>>>>>> of connection timeout.  I can telnet to the port, so
>>>>>>>> maybe the timeout is too short?  Or I'm missing some config
>>>>>>>> I should have done for the client samples?  I'm going by
the
>>>>>>>> README at:
>>>>>>>> https://github.com/SciGaP/Airavata-PHP-Client-Samples/blob/master/client-samples/README-PHP-CLI-Examples.txt
>>>>>>>> 
>>>>>>>> and getting this kind of error:
>>>>>>>> 
>>>>>>>> john@buffy:/media/mdhpss/scigap/phpclientsamples/Airavata-PHP-Client-Samples/client-samples$
telnet gw111.iu.xsede.org 8930
>>>>>>>> Trying 149.165.228.109...
>>>>>>>> Connected to gw111.iu.xsede.org.
>>>>>>>> Escape character is '^]'.
>>>>>>>> ^]
>>>>>>>> 
>>>>>>>> telnet> quit
>>>>>>>> Connection closed.
>>>>>>>> john@buffy:/media/mdhpss/scigap/phpclientsamples/Airavata-PHP-Client-Samples/client-samples$
php createProject.php kenneth
>>>>>>>> PHP Fatal error:  Uncaught exception 'Thrift\Exception\TException'
with message 'TSocket: Could not connect to gw111.iu.xsede.org:8930 (Connection timed out
[110])' in /media/mdhpss/scigap/phpclientsamples/Airavata-PHP-Client-Samples/lib/Thrift/Transport/TSocket.php:235
>>>>>>>> Stack trace:
>>>>>>>> #0 /media/mdhpss/scigap/phpclientsamples/Airavata-PHP-Client-Samples/client-samples/createProject.php(45):
Thrift\Transport\TSocket->open()
>>>>>>>> #1 {main}
>>>>>>>> thrown in /media/mdhpss/scigap/phpclientsamples/Airavata-PHP-Client-Samples/lib/Thrift/Transport/TSocket.php
on line 235
>>>>>>>> john@buffy:/media/mdhpss/scigap/phpclientsamples/Airavata-PHP-Client-Samples/client-samples$
>>>>>>>> 
>>>>>>>> 
>>>>>>>> On Wed, Jun 25, 2014 at 02:56:22PM -0400, Marlon Pierce wrote:
>>>>>>>>> The CCGrid demo server is running on gw111.iu.xsede.org.
This where
>>>>>>>>> the PHP client samples from GitHub are pointing.
>>>>>>>>> 
>>>>>>>>> Marlon
>>>>>>>>> 
>>>>>>>>> On 6/25/14, 2:28 PM, K Yoshimoto wrote:
>>>>>>>>>> Is there an instance of SciGap services available
for testing
>>>>>>>>>> and development now?
>>>>>>>>>> 
>>>>>>>>>> On Wed, Jun 25, 2014 at 01:09:58PM -0400, Suresh
Marru wrote:
>>>>>>>>>>> Hi Kenneth,
>>>>>>>>>>> 
>>>>>>>>>>> I see that Marlon already listed the PHP samples
below - https://github.com/SciGaP/PHP-Reference-Gateway
>>>>>>>>>>> 
>>>>>>>>>>> The functionality which is not so easy is the
registration of the applications.
>>>>>>>>>>> 
>>>>>>>>>>> Suresh
>>>>>>>>>>> 
>>>>>>>>>>> 
>>>>>>>>>>> On Jun 23, 2014, at 12:58 PM, Marlon Pierce <marpierc@iu.edu>
wrote:
>>>>>>>>>>> 
>>>>>>>>>>>> Hi Kenneth--
>>>>>>>>>>>> 
>>>>>>>>>>>> The Airavata repo information is here: http://airavata.apache.org/development/source.html.
 This includes a link to the GitHub mirror, which is probably best.
>>>>>>>>>>>> 
>>>>>>>>>>>> The Thrift definitions are at https://github.com/apache/airavata/tree/master/airavata-api/thrift-interface-descriptions.
Our apachecon slides may be useful supplements: http://www.slideshare.net/smarru/restless-design-with-apache-thrift-experiances
>>>>>>>>>>>> 
>>>>>>>>>>>> The PHP command line samples (https://github.com/SciGaP/Airavata-PHP-Client-Samples/),
not currently in Apache, illustrate how the API works.
>>>>>>>>>>>> 
>>>>>>>>>>>> Marlon
>>>>>>>>>>>> 
>>>>>>>>>>>> On 6/23/14, 12:34 PM, K Yoshimoto wrote:
>>>>>>>>>>>>> Suresh,
>>>>>>>>>>>>> 
>>>>>>>>>>>>> Could you please send out a URL to the
current Airavata repo?
>>>>>>>>>>>>> And the API documentation?  I want to
look through the current
>>>>>>>>>>>>> interface code.
>>>>>>>>>>>>> 
>>>>>>>>>>>>> Kenneth
>>>>>>>>>>>>> 
>>>>>>>>>>>>> On Mon, Jun 23, 2014 at 08:13:14AM -0400,
Suresh Marru wrote:
>>>>>>>>>>>>>> Hi Borries,
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> Can we take up on your demo offer
and make an additional request to you and Gary?  Please bare my long email, I am sorry I do
not have much time to make this shorter.
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> XSEDE 14 tutorial is our opportunity
to mint more and more Demeler's, Miller's, Majumdhar's and Pampidighantam's. I say this seriously,
all of you have sacrificed your own science and enabled lot of more science by 100's of fellow
researchers. There are not many more who can dedicate their careers to do this. So we need
to show building and operating gateways will not necessarily suck up their own science. SciGaP
will absorb the banal tasks and help them focus on their gateway mission - "server a community
of users enabling better science".
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> One way we can demonstrate SciGaP
is by hands-on tutorial of building a bare-bone gateway within half a day. But in addition,
we need to assure its beyond a tutorial and is indeed the same software used by partnering
gateways. We need a beta version of LIMS instance which uses same thrift API's and served
by the same set of Airavata services. Preciously put, it will be nice to say current ultrascan
production using Airavata 0.8, but there is also development in progress to roll out Ultrascan
with latest Airavata (0.12).
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> Can Raminder and Gary work together
to make this happen within the next 3 weeks?
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> Hi All,
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> As for the distracting banter, if
you all prefer to know the details, there is nothing secretive about it. We got a highest
panel rating standing ahead of 176 other proposals. Its natural the community wants us to
succeed. Some say this by clapping hands, some challenge us by saying we will fail. Both are
equal motivators. I am personally not so worried about iPlant. But I am a bit concerned the
negativity is spreading and we directly took punches from Jay Boisseau, Dan Stanzione, Shantenu
Jha and unfortunately from Nancy Wilkins-Diehr (on SciGaP and S2I2 gateway institute context,
which we will resolve very soon). There are other folks too, but there is nothing we can or
should (in my opinion) do about this directly. Its simply not worth and is a unnecessary drain
of pre-XSEDE 14 energy.
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> We certainly need to disprove "every
one in this team will do their own thing". But I do not see SciGaP as Airavata+CIPRES+NSG+Ultrascan
nor IU+SDSC+UTHSA. To me, SciGaP means SciGaP, the mission is to provides software and centralized
operations and the team believes and practices full openness - both software and operations.
We wrote a proposal on these beliefs and we need to play our part in changing the gateway
eco system working alongside friends and some more dear friends.
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> Suresh
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> On Jun 21, 2014, at 10:34 PM, Marlon
Pierce <marpierc@iu.edu> wrote:
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> Hi Borries--
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> I was a little too frustrated
on Friday in dealing with the iPlant Agave lead developer.  Mark and Terri have had previous
bad experiences. It isn't worth any time or effort, I believe.  Working to succeed in our
SSI is the best path.
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> Marlon
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> On 6/21/14, 9:10 PM, Miller,
Mark wrote:
>>>>>>>>>>>>>>>> I am not at all surprised,
we tried to interact with those people and were told in no uncertain terms we should just
discard our software and wait for further instructions. This was not a suggestion we were
comfortable with. There are plenty of ways we can interact productively with any given infrastructure.
>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>> In my experience there are
always groups who say negative things as a tool for positioning their project with a competitive
advantage, and to sew panic among collaborators.
>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>> I tend to be driven by what
makes sense for users, and ignore the political winds from people who are not stakeholders.
>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>> However, it certainly would
be useful to learn more about what this conversation was about, as this is a pretty strong
statement.
>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>> Mark
>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>> On Jun 21, 2014, at 3:51
AM, "Borries Demeler" <demeler@biochem.uthscsa.edu> wrote:
>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>> Hi Marlon,
>>>>>>>>>>>>>>>>> I am a bit alarmed by
your statement that apparently there are
>>>>>>>>>>>>>>>>> folks, esp. at Austin,
who expect and want us to fail. I would
>>>>>>>>>>>>>>>>> appreciate if you could
elaborate on what exactly is bugging
>>>>>>>>>>>>>>>>> these people so maybe
we could resolve this? We cannot afford
>>>>>>>>>>>>>>>>> to have this antipathy
against our project, and I want to prove
>>>>>>>>>>>>>>>>> them wrong. It is absolutely
clear that at least the UltraScan
>>>>>>>>>>>>>>>>> integration has been
a big success and can serve as a model
>>>>>>>>>>>>>>>>> for other integrations.
I am sure with CIPRES similar success
>>>>>>>>>>>>>>>>> can be achieved, but
I may need some more background info to
>>>>>>>>>>>>>>>>> understand exactly what
issues are raising concern.
>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>> A certain amount of re-write
is always necessary, and we fully
>>>>>>>>>>>>>>>>> expected and welcomed
the opportunities to improve our software,
>>>>>>>>>>>>>>>>> and there is a lot of
code we ultimately abandoned, re-wrote
>>>>>>>>>>>>>>>>> from scratch or refactored/adapted
to new and better methods.
>>>>>>>>>>>>>>>>> I must honestly say that
I feel very fortunate that we have
>>>>>>>>>>>>>>>>> your group and the AIravata
concept to make our software more
>>>>>>>>>>>>>>>>> portable and robust.
These are the two major benefits for us,
>>>>>>>>>>>>>>>>> but there are others,
too.
>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>> Any chance we can give
these people a demo of UltraScan to
>>>>>>>>>>>>>>>>> show them how well it
worked?
>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>> -b.
>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>> On Fri, Jun 20, 2014
at 09:32:18PM -0400, Marlon Pierce wrote:
>>>>>>>>>>>>>>>>>> Thanks for your thoughts
on the matter, Gary.  I intend to give a
>>>>>>>>>>>>>>>>>> follow up in more
detail on Monday.  In any event, we will need a
>>>>>>>>>>>>>>>>>> path forward to deliver
on the project goals.  I was reminded
>>>>>>>>>>>>>>>>>> unpleasantly and
unambiguously today that at least one of our
>>>>>>>>>>>>>>>>>> colleagues in Austin
wants and expects us to fail.
>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>> Marlon
>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> On 6/20/14, 9:25
PM, Gary E. Gorbet wrote:
>>>>>>>>>>>>>>>>>>> I was interested
in the presentation by Saminda on CIPRES-Airvata integration. I would like to comment on one
issue that arose.
>>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> The presentation
spoke of a "new CIPRES Portal". The CIPRES developers expressed dismay at the prospect of
a "total rewrite" of CIPRES to make it compatible with Airavata. But I did not get the impression
that a complete rewrite was being proposed. Rather, I interpret "CIPRES Portal" as that portion
of the Gateway that is directly concerned with interface with the middleware. Terri expressed
a desire to limit changes to the CIPRES SDK. I see no conflict to that idea in what Saminda
proposed.
>>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> I have only a
cursory understanding of CIPRES. However, I believe that its functions broadly mirror those
of the UltraScan gateway that I work on. The UltraScan gateway consists primarily of a set
of PHP scripts. There are also CSS and JS (JavaScript) components. I would not anticipate
making any changes to CSS or JS as the gateway transitions to SciGaP, since they are concerned
with web page presentation and parameter specification. Nor would I think that many of the
PHP scripts would change.
>>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> Where I do see
a necessity for a rewrite is in specific scripts that directly interface to the middleware.
Two examples come to mind.
>>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> First, there
is the basic action at submit time. Presently, task parameters are incapsulated into an XML
job file. Analysis parameters are incapsulated into another XML file that is combined with
data files in a single archive (TAR) file. The task XML and archive files are passed to GFAC
in the course of posting to the GFAC service via URI. In SciGaP, this procedure will be changed
considerably. The post to a service will be replaced with Thrift API calls. The task will
be defined by building a structure that is passed via Thrift. The changes are substantial
enough that it makes sense to me to rewrite the specific PHP script charged with submitting
a job.
>>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> Similarly, the
job monitoring procedure will change considerably with SciGaP. Where currently UltraScan and
GFAC communicate task status via a shared database, SciGaP will put all responsibility for
determining task status on Airavata/SciGaP and the gateway will be advised of changes by a
callback mechanism. The gateway will advise that status may need re-evaluation with Thrift
calls. So, here again, the specific parts of the code related to task state monitoring should
probably be rewritten.
>>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> Perhaps CIPRES
has similar functions and modules. As I said, I don't know the specifics of CIPRES, but I
imagine that user web interface code need not be changed, while parts of the code dealing
with middleware interface might need substantial enough changes that a rewrite of selected
modules would be indicated. Perhaps all that code is within CIPRES's SDK.
>>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> Saminda or Marlon:
please correct me if I got this wrong. My impression was the "total rewrite" might better
be characterized as "selective modular rewrite".
>>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> - Gary
>>>>>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>>>>>> To unsubscribe
from this group and stop receiving emails from it, send an email to dev+unsubscribe@scigap.org.
>>>>>>>>>>>>>>>>>> To unsubscribe from
this group and stop receiving emails from it, send an email to dev+unsubscribe@scigap.org.
>>>>>>>>>>>>>>>>> To unsubscribe from this
group and stop receiving emails from it, send an email to dev+unsubscribe@scigap.org.
>>>>>>>>>>>>>>> To unsubscribe from this group
and stop receiving emails from it, send an email to dev+unsubscribe@scigap.org.
>>>>>>>>>>>>>> To unsubscribe from this group and
stop receiving emails from it, send an email to dev+unsubscribe@scigap.org.
>>>>>>>>>>>>> To unsubscribe from this group and stop
receiving emails from it, send an email to dev+unsubscribe@scigap.org.
>>>>>>>>>>>> To unsubscribe from this group and stop receiving
emails from it, send an email to dev+unsubscribe@scigap.org.
>>>>>>>>>>> To unsubscribe from this group and stop receiving
emails from it, send an email to dev+unsubscribe@scigap.org.
>>>>>>> To unsubscribe from this group and stop receiving emails from
it, send an email to dev+unsubscribe@scigap.org.
>>>>>> To unsubscribe from this group and stop receiving emails from it,
send an email to dev+unsubscribe@scigap.org.
>>>>> To unsubscribe from this group and stop receiving emails from it, send
an email to dev+unsubscribe@scigap.org.
>>>> To unsubscribe from this group and stop receiving emails from it, send an
email to dev+unsubscribe@scigap.org.
>> To unsubscribe from this group and stop receiving emails from it, send an email to
dev+unsubscribe@scigap.org.

----- End forwarded message -----

Mime
View raw message