PTB Meeting

Europe/Zurich
Teleconference

Teleconference

Description
PTB meeting, 14:00-16:00, 30 November, 2010 Calling details: skype test conferencing service in Lund (call mailmagnes in skype) Screensharing: https://my.dimdim.com/arc-emi (12345) Proposed agenda: ---------------- - Approval of previous meeting notes, follow up on actions - Follow up of the Prague all-hands - EMI "UI", clients for EMI products (see mail from Danilo) - missing glite components from the DNA1.3.1 (see Maria's mail) - AoB
PTB meeting, 14:00-16:05, 30 November, 2010

Present: Laurence, Morris, Marco,  Balazs (notes),  Francesco (invited)
Excused: Patrick, John


1)  Unclear (glite) components (PT, Partner)

On the request of Maria Alandes Pradillo the PTB scheduled the discussion of certain (mostly) glite production components those "EMI status" was unclear.  The PTB wanted to resolve the uncertainty around these components by clearly identifying the responsible partner and product team in case the components are decided to be part of EMI software stack.  The unanimous decisions are the following:

- lcg-info-dynamic-scheduler-generic:  part of EMI, assigned to NIKHEF and glite job management PT

- lcg-info-dynamic-software: part of EMI, assigned to CERN and glite job management PT

- lcg-info-provider-software: not part of EMI, obsoleted code.

- lcg-tags and lcg-ManageVOTags: part of EMI, assigned to CERN, glite job management PT

- Batch system bindings including both the submission and  infosys parts:

 --Torgue: INFN (sub) Nikhef (info)
 --SGE: CESGA (both sub and info)
 --LSF: INFN (sub), CERN (info)
 --Condor: no EMI partner behind

All of the above modules are part of EMI and taken care of the respective partners and belong to the glite job management PT.

- GENERAL rule about info providers for EMI services (e.g. dcache, DPM, FTS, A-REX, etc..): Every service is responsible to write and support its own info provider unless it is explicitly stated otherwise.

- glite-CLUSTER: this node type (metapackage) is out of scope of EMI. It is not applicable to CREAM.


2) node types, installation metapackages

Long discussion about the role of deployment metapackages such as "UI" "WN". These are not considered as EMI components, no product teams are responsible for them.

However, creation of certain metapackages will be necessary, the work is considered as a kind of combination of "integration" (JRA1) and software packaging/distribution/deployment (SA1) area. Details to be discussed later, once we get closer to EMI-1 code freeze and release dates.

The PTB considered two specfic node types to see a short/middle term possibilities and steps.

a) EMI-UI: General consensus over the need for a EMI-UI box, already on the short term. This "box" should contain as much as possible from the current clients of the four mws. PTB decided the creation of a new task force within JRA1:

EMI-UI task force:
*mandate:  define content (list commands, utilities) of an UI box, define a metapackage, work on integration, create a user interface box, work on config, documentation. Main goal is to integrate the existing clients as much as possible.
*members: every mw to delegate one person
*leader: to be found
*timeline: a EMI-UI box should be ready by EMI-I

b) WN:
The worker node metapackage is required for glite sites. EMI-I release must offer such a deployment concept. CERN is responsible to provide the WN metapackage for EMI-I, mostly (if not exclusively) based on glite components. Later, after EMI-I, the entire concept of WN software will be revisited to see how it fits with other mws.


3) Platform

Since Francesco (SA1) was present, therefore the PTB took the opportunity to discuss little bit  "EMI platforms". It was generally shared that before list of mandatory platforms are decided it is worth to survey the PTs to find out how much effort is needed on their side to port/make their code available on a specific platform.


4) Follow up of the Prague all-hands meeting.

- data area: no urgent actions, according to the data summary talk by Patrick (was sent by mail)

- compute area: EMI-ES interface group needs to deliver the "agreement" still this year. It is on track, see the group's progress, latest draft and their tight schedule at
https://twiki.cern.ch/twiki/bin/view/EMI/EmiExecutionService

- security area: delegation task force was launched with all the relevant teams involved

- infra area:
 -- monitoring task force urgently need person/partner assigned. It is curently an empty task force though EGI-Inspire expects this task force to deliver at least the Nagios probes as part of EMI-I
-- Cloud and Virtualization task forces: agreed not to merge the two groups yet. PTB waits with the decision.
-- service registry PT: progressing well, early design is expected soon
-- info area harmonization:  work starts soon
- service management task force: after the all-hands, the TF is in good hands with Jason. The directions are clarified.

5) AOB

- next meetings: 14.12, 11.01 (booked in indico)

- meeting notes of 9.11.2010 got approved. There are no pending action items.



ACTIONS:
--------
- Balazs initiates a survey to find out the work needed to support a list of specific platforms
- Balazs: reports the situation with the monitoring task force to the PEB
- Balazs: schedules the "overview of EMI aggrements" topic for the the next PTB meeting
- Balazs: makes sure the "components" decisions are properly publicized and information reaches partners/PTs.

There are minutes attached to this event. Show them.
The agenda of this meeting is empty