PPS Pilot Follow-up Meeting Minutes Fri 29 Mai 2009
- Date: Fri 29 May 2009
- Agenda: 59829
- Description: Pilot of Cream CE: check-point
- Chair: Antonio Retico
Attendance
- PPS: Antonio Retico
- CERN: Apologies
- FZK: Apologies
- PIC: Raquel Munoz, Oscar Olivier
- CNAF: Absent
- CMS: Absent
- Alice: Apologies
- JRA1/Cream/WMS: Massimo Sgaravatto, Sara Bertocco
- SA3: Gianni Pucciani
Review of action items (tasks)
Not covered.
Status and results of the pilot service (by VOs and sites)
Acceptance test by SA3
Gianni: no progress to report. People from GRNET during the weekly phone conference was given other priorities so they have postponed the task of running those tests.
Antonio: this is concerning. Result from this activity were expectend by mid May. It will be escalated to Nick (responsible for the definition of the acceptance criteria)
Note: after the meeting Antonio and Nick talked to Oliver. two actions were taken (1. a request for an estimation was sent to GRNET 2. a proposal to get involved was sent to the Russian site which already collaborated with Laurence in the benchmarking activity described at the EGEE User Forum)
CMS
no report from CMS.
Antonio: I have the impression that we are at an impasse here. What we would really like to see is CMS to use the WMS at CNAf as part of their steady-state production and not as an artificial test (so put in round-robin with the other WMSs). This configuration is impossible to do with the current version of the production WMS becasue this would cause jobs sent to
CREAM via standard WMS to fail.
Unfortunately the deployment of the WMS 3.2 (
PATCH:2597) currently in pre-production and including the new version of ICE cannot be expected very soon and even if it went out today hardly sites would upgrade because the concurrent STEP09.
Massimo: the content of
PATCH:2459 (obsoleted) is the same version of ICE deployed with
PATCH:2597 . Cna't we deploy this one?
Antonio: this is a good idea and actually could make the deployment of ICE easier. Gianni and I will invvestigate with Andreas if this can be done. We could set this patch directly to "certified" and include it directly in the next batch for production. That would allow CMS eventually to request the sites providing their WMS to upgrade if they want to use
CREAM.
Note: Andrea Sciaba' joined when the meeting was closed and was adjourned on what was said. He thinks that a test with a secondary instance of CRAB still makes sense to be done (in fact CMAS has never done such a test since the beginning of the pilot). It is up to Stefano Lacaprara to arrange such a set-up. Andrea remarked that the sites offering support for CMS in production should satisfy the well known pre-conditions for CMS to run (afs mounted on WNs with SW area visible and close SE pointing to a SE containing CMS data sets). Antonio will make a survey between the sites publishing support to CMS in the info system to verify that they satisfy these conditions as well.
PIC
Antonio: the two
CREAM at PIC can be regarded as production machines at all effects. If the site want they could be even registered in the GOCDB as such. For the pilot specific activity instead, it would be very useful if the condor based cream installation managed by Marc Rodriguez could be opened to receive the tests from SA3, because that would cover a particular use case.
Oscar: there are not moany WNs connected to this instance.
Antonio: the test is focused on the
CREAM node rather than on the WN on the back-end. In fact even only with two WN the test could be performed, the only condition is that the
CREAM machine suits the specifications.
Status and results of the development (by developers)
Massimo: for
CREAM we are currently working on
PATCH:2666 that fixes several bugs. The we worked on the
BUG:47911. We were able to track the cause of a huge memory consumption which now is hopefully fixed. The fix is still under test
Antonio: does the WMS at CNAf conatin this last fix?
Massimo: No the WMS at CNAF is for the time being at the equivalent of
PATCH:2459
Open Issues (by VOs, sites, deployment teams)
None
Recommendations for release and deployment
None
Decision about termination/extension of the pilot
the next check point is fixed for Tuesday 16th of June at 10.00
AOB