November 2005 - cddlm-wg

Last call for Application Contents Service Specification 1.0
by Keisuke Fukui 08 Mar '06

08 Mar '06

Folks in ogsa-wg, cddlm-wg and acs-wg, We, acs-wg, has come to completion of the ACS Specification 1.0 after discussions and reviews. This is the last call for comments before submission to the GGF Editor for publication. If you have comments, please get them to me by 9 am Japan time on Monday November 7. If there are no significant comments I plan to submit the document on Monday afternoon. Below is the link for the document: https://forge.gridforum.org/projects/acs-wg/document/draft-ggf-acs-spec-1.0… Regards, Keisuke Fukui ACS-WG co-chair

1 1

Doubts about SWITCH
by Flávio Roberto Santos / Projeto Ourgrid 30 Nov '05

30 Nov '05

Hi all, We have been trying to implement the Switch flow control and we've got some doubts: 1. In the 'component-model.xsd', the Switch element has the following element: <xsd:element ref="case" minOccurs="0" maxOccurs="unbounded"/>. Is the 'minOccurs' value ZERO correct? I think that this attribute should have the value 1 (one), unless there will exist an "otherwise" without a "case" element. 2. Is it possible to have more than one cmp:switch element in the same level of a CDL and referring to the same lifecycle? If so, does the order of the switch matter? 3. Should all the components execute a lifecycle operation or a switch can avoid a component to be called in a certain operation? 4. How to proceed when there is more than one ref to the same component? Should we execute more than one time the same operation on that component? 5. As we can have more than one "case" element for the swith, what does the otherwise mean? Should it execute when all case conditions are false or when the last one is false? Thanks in advance, Flavio and Ayla -- Flavio Roberto Santos Undergraduate student at Universidade Federal de Campina Grande OurGrid Team Member - www.ourgrid.org My webpage - http://www.ourgrid.org/~flavio -- "As far as the laws of mathematics refer to reality, they are not certain, and as far as they are certain, they do not refer to reality." -- Albert Einstein

1 0

Meeting tomorrow, 11/30, 6am PST
by Milojicic, Dejan S (HP Labs) 30 Nov '05

30 Nov '05

Hi, There will be our usual meeting tomorrow. The following is tentative agenda, please let me know if you want to address anything else: 1. Update on the testing 2. Continue planning next GGF (hotel reservations, what do we plan to do at GGF, etc.) 3. Continue discussions on interoperability. Please use our usual phone numbers: +1 866-639-4732 (toll free while in US) Code 2362906 (followed by a '#' sign) +1 574-948-0379 (toll & international) Thanks, Dejan.

1 0

Re: [cddlm] status on the CDL Test environment
by Steve Loughran 24 Nov '05

24 Nov '05

Jun Tatemura wrote: > Steve, > I finally added the file to the CVS. Sorry for the delay. no problem, it wasnt on my critical path. I knew when it checked in as gump sent me email: http://vmgump.apache.org/gump/public/ggf-cddlm/cddlm/index.html I will have to tweak the build so that it skips java1.5+ stuff on a 1.4 system, which is what vmgump is. > Please also note that sync between developer CVS and anonymous CVS has been > disabled since Nov 11 accooding to SourceForge.net Site status page. > sourceforge are being very unreliable w/ anoncvs right now. Apparently all projects beginning with "J" have had anoncvs broken for a few weeks. But, since gump picked up the changes (it has anonymous access), all must be recovering. FYI, gump is the daily clean-build-all-OSS-java projects that apache runs; you declare your dependencies and it builds things in the right order, complaining when things fail. the nightly build we've been running for cddlm verifies the test documents against the example schemas, so we can verify that our XSD is valid as far as we think, and our test docs are valid against the spec. even without the java1.5 compile, this is something we should extend to new test docs. I'm not actually building my CDL deployapi modules under gump because it is java1.4 only; if/when java1.5 comes back we can do that. thanks, -steve

1 0

status on the CDL Test environment
by Jun Tatemura 24 Nov '05

24 Nov '05

This is a preview of the CDL test cases and testing environment. A CDL implementor can (1) implement these Java interfaces and use this testing environment, or (2) implement another testing environment that processes test cases in XML. For (1), you need to implement CDLProcessor and CDLProcessorFactory. Please also look at Tester if you want to see how they are called in detail. I will give you a complete test environment with more test cases by the meeting but please check these files (XML test case and Java interfaces) to see they are okay to you. Thanks, Jun

2 2

test docs
by Steve Loughran 16 Nov '05

16 Nov '05

Stuart, I've moved the api test doc under doc/test , and split out one for system testing into the same directory. Both are .doc format. And the api doc has been updated with changes discussed on the phone, plus some ones I did yesterday. I'm cheating and using header numbering to number my tests. This is cheating as when stuff moves around, they change. We really need a proper naming scheme. -steve

1 0

RE: Meeting tomorrow, 11/16, 6am PST
by Milojicic, Dejan S 16 Nov '05

16 Nov '05

Hi, > 1. Update on the testing Stuart's feedback to Jun, same as Steve's separate valid tests from invalid ones (those that should fail); Stuart would leave the CDL wrappers in his tests. Steve: needs to put it in the CVS. Stuart's comments on Steve's work: PDF is hard, get some editable document (Steve: it is in CVS). #2, there are a lot of systems tests. it would be useful to have a separate systems document. Steve: it goes hand in hand, functional and systems tests. Steve: we can split notifications in a separate testing, but the rest is blended together. Stuart: testapi-3, page 2, create a system and destroy it; step 5 should be go back to portal and see if the system still exists. Steve ok's this one. Page 5, the paired portal test, it is a systems test, not sure if he can do this; should move to systems test. Steve will pull out the systems tests and Stuart will verify. Stuart will come back with his tests maybe even earlier than 2 weeks after Steve's. Dejan: Are we going to have a common documents for tests. Stuart volunteered to do this, the only obstacle at the moment is that Jun's documents are just code right now. Jun agreed to have the output from XML and pass it to Stuart. > 2. Update from the OGSA meeting last week Stuart and Dejan attended, it went fine, we addressed Andrews Grimshaw's questions, we also started discussions with others. We will have a separate, offline discussions with Jay on where to cut the tree between CDDLM and provisioning. > 3. Plan meeting with Jay et al. Dejan will invite people designated to attend, will cc CDDLM all CDDLM members are welcome (and encouraged) to join in the discussions. > 4. Continue planning next GGF (hotel reservations, what do we > plan to do at GGF, etc.) Please be reminded that one day after regular GGF, there will be OGSA F2F, that everyone should plan on attending. > 5. Continue discussions on interoperability. How much do we open up the process? This is something to ponder. There will be NO meeting next week. Next meeting will be on 11/30. Thanks, Dejan.

1 0

Meeting tomorrow, 11/16, 6am PST
by Milojicic, Dejan S 16 Nov '05

16 Nov '05

Hi, There will be our usual meeting tomorrow. The following is tentative agenda, please let me know if you want to address anything else: 1. Update on the testing 2. Update from the OGSA meeting last week 3. Plan meeting with Jay et al. 4. Continue planning next GGF (hotel reservations, what do we plan to do at GGF, etc.) 5. Continue discussions on interoperability. Please use our usual phone numbers: +1 866-639-4732 (toll free while in US) Code 2362906 (followed by a '#' sign) +1 574-948-0379 (toll & international) Thanks, Dejan.

1 0

deployment test plan
by Steve Loughran 14 Nov '05

14 Nov '05

Here is an updated draft of the deployment test plan. It is incomplete, but I dont intend to make any enhancements to it for a while, as I have to go and build things. The most incompletely specified bit is the WS-N section, which is a metric of how complete my implementation of that is. If anyone is ahead of me there and wants to write those tests, go ahead. We also need to think of how to test the entire system, with tests for component model+deployAPI. -steve

1 0

Testing parsing with the deployment API
by Steve Loughran 14 Nov '05

14 Nov '05

I've been thinking about how to add a testing operation to the deployment API; a way to get at the fully parsed document. How about we have the option of a new property in a new namespace (xmlns:diagnostics="...") :- diagnostics:ParsedDescriptor If implemented, this read only resource property would contain the entirety of the parsed document, set during the initialize() operation. Implementations must list the qname for this in their ManagementCapabilities operation, plus, if they want to be OGSA-BPI compliant, in that thing that that spec asks for. The nice thing about this approach is that as well as adding no new operations to the API, it lets anyone working with a deployment find out what it was that was actually instantiated. So it will benefit users as well as CDL implementors. Oh, and it needs no extra WSDL, just some XSD work. -steve

1 0