PPS Pilot Follow-up Meeting Minutes Tue 01 Jul 2008

• Date: Tue 01 Jul 2008
• Agenda: 36931
• Description: pilot of Cream CE: check-point
• Chair: Antonio Retico

Attendance

• PPS: Antonio Retico, Nick Thackray (joined later)
• CNAF: Daniele Cesini, Danilo Dongiovanni
• FZK: Angela Poschlad
• UPATRAS: Vasiliios Kolonis, George Goulas
• PIC: -
• Cream (Cluster of Competence): Massimo Sgaravatto
• SAM: represented by Antonio
• Nagios: -
• CMS: Andrea Sciaba', Enzo Miccio (joined later)
• Alice: Patricia Mendez (joined later)
• Atlas: -
• LHCb: Roberto Santinelli (joined later)

Status of the pilot service (by VOs and sites)

progresses on tasks

- status of the Pilot (phase1)

The status of the subtasks of TASK:7143 was reviewed

Notes:

• TASK:7121 : done
• TASK:7139 : updated with request to configure VO SW dir (see later among CMS comments)
• TASK:7140 : done. a new TASK:7274 was opened for 3 weeks of service management
• TASK:7153 TASK:7144: waiting on SAM installation for PPS to be completed: currently bdii2oracle and publishing webservice are missing. TASK:7275 is created for that
• TASK:

Initial planning of phase2

Massimo pointed out that a better version of the ICE-enabled WMS is currently in phase of pre-certification (PATCH:1841). JRA1 would prefer this one to be the version used for phase 2. The ICE-enabled WMS we have now works, but it suffers for a number of bugs and would not sustain stress testing very well.

Furthermore Sara is preparing a new version of the yaim function to fix the problem found during phase one with glexec

---

An update from Massimo was received after the meeting

-----Original Message-----
From: Massimo Sgaravatto - INFN Padova [mailto:massimo.sgaravatto@pd.infn.it] 
Sent: Wednesday, July 02, 2008 9:24 AM
To: Massimo Sgaravatto
Cc: egee-pps-pilot-cream; Sara Bertocco
Subject: Re: A clarification on CREAM installation procedure: the glexec issue

Sara has fixed the yaim installation procedure 
(the new one is implemented by glite-yaim-cream-ce-4.0.4-10).
You can check this patch:

 

to see the list of bugs that have been fixed


At the usual link:

http://igrelease.forge.cnaf.infn.it/doku.php?id=doc:guides:devel:install-cream31-devel

you can find the installation/update instructions

And as usual you feedback will be highly appreciated

					Cheers, Massimo

---

Daniele: should we be publishing the services in the PPS BDII?

Antonio: we definitely should, because this is needed for the SAM test case

George and Vasilios asked when they will be involved for phase2.

Antonio: phase 2 was supposed to start today (1st of July). The second phase is about installing the cream CEs in the production environment. The best contribution that UPATRAS could give now wold be to deploy to creamCE in the production environment (using the same PPS machines) as soon as the new YAIM procedure is available (now it is, with glite-yaim-cream-ce-4.0.4-10)

George: we are supposed to work strictly in PPS, so we need to verify if we can publish services in production

Antonio: In theory, if this is strtictly required by the experiments we could install some Cream services and move one or both the WMS we have now in production. However, in consideration of what Massimo and George said, it could be worth waiting for the software to settle down a bit more, provided that the VOs have way to make some use of the current set-up.

The up-2 date status of the pilot is in PpsPilotCream

Involvement of experiments during phase1

Antonio made the same qestions to the three VOs attending

1. How is your experiment planning to submit to cream CE? (via WMS, direct submission...)
2. Is the current set-up sufficient for the experiment to start testing or we need to move towards the production system?
3. When so you plan to start your testing activity

Alice

1. Alice is going to submit via the WMS.
2. The environment has to be consistent, but Alice can run tests both in the PPS or production environment. The only requirement is a dedicated VOBOX to be set-up at the site.
   That's why Patricia suggested an installation of the CreamCE at CERN, because they have already a second VOBOX to use here.
   Angela proposed in alternative to set up a dedicated VOBOX at FZK-PPS. This cannot be a very powerful machine though.
   Angela's proposal was accepted by Patricia
3. Alice will start as soon as it will receive communication (by Antonio) that the installation of the VOBOX is completed

CMS

1. CMS is going to use the WMS and they agree with Massimo's idea of delaying the deployment to production to when the WMS will be in a steadier condition
2. A simple test that could be done with this set-up is to tweak the SAM client and try and submit functional tests to the Cream CE. It is mandator for that to have the VO SW area visible to the WNs. It is possible to use the PPS infrasturcture as is now, but the CREAM-CE service has to be recognised in the SAM DB.
   

Antonio

the SAM server in PPS is supposed to be covering this use case but it is not ready yet (a TASK:7275 has been created to monitor the progress).

1. CMS will start working as soon as the SAM service will be available in PPS (Antonio to notify).

2 additional questions by CMS:

1. When the WMS will be ready?
   • Massimo: the certification process has changed in EGEEIII. Now the WMS is in phase of pre-certification, joint activity done by SA1 and JRA1. After this phase is over the standard formal certification should start, which is supposed to take less time because most of the issues will have been found earlier in the process. Impossible to have a definite answer though
2. What are the next steps for the deployment of Cream?
   • No reply iven during the meeting. As a partial reply I post here an excerpt of the SA3 certification guidelines, as drafted by Oliver
     

     Comment - we are *not* [yet] certifying ICE, and would like to avoid
     difficulties in interpreting test results if possible. Di will make a
     judgement as to whether we can simply loop over the CREAM CLI tests we
     will have in order to do the appropriate scalability validation. If so,
     this is the approach we will take. If not, we will take the ICE rpms and
     upgrade the testbed WMS. There's is a question over testing
     proxy-renewal if we don't use the WMS.
     
     When CREAM is released to production, it will be advertised as being
     made available for larger sites to install in parallel with their
     existing lcg-CEs, not as a replacement for them. In this way we will
     soon have a pool of CREAM CEs exposed to production work patterns and
     loading, without endangering availability of resources. 
           

LHCb

LHCb haas been deeply involved in beta-testing of Cream two years ago and it is now implementing job submission using "ortogonal" methods. Therefore , for the time being, LHCb is not interested in taking part into this activity.

However they asked for an updated version of the user guide in order to explore the possibility in the future to take advantage of cream features within their submission model.

Recommendations for release and deployment

VO SW area to be visible on the WNs

VOBOX for Alice needed to test

SAM service must be set-up for CMS testing

Decision about termination/extension of the pilot

In consideration of what discussed above, the decision is made to extend this phase of the pilot within the PPS infrastructure and to delay the deployment in production of three weeks from now.

The next check-point meeitng will be on the 22nd of July

AOB

Actions

See child tasks of https://savannah.cern.ch/task/?7143

---