CREAM CE has been released to production, but will not yet be selected by default WMS match-making.
Next week, many gLite 3.0 services will be made obsolete.
Attendance
The meeting started at 16:04. We should really try to start at 16:00 out of respect to those who make the effort to connect on time.
The list of attendees is taken directly from web conference display. Please ensure that you open a browser window and enter your name and affiliation - even if you also dial in by phone!
EGEE & WLCG
OCC / CERN ROC: John Shade, Antonio Retico, Nick Thackray, Steve Traylen, Diana Bosio
Issues mainly from AP region, some UK, and one at CERN.
Does ROC UKI receive notifications from GGUS? Jeremy will follow-up (wasn't aware of any problems).
Russia will follow-up the APEL problem on one of their sites
Steve will personally follow-up with AP ROC
PPS Reports
CREAM CE pilot is entering a new phase. The plan is to setup 20 CEs, 3 Italian sites and 1 German site currently involved, others would be welcome.
PPS received two releases this week (mainly removal of obsolete packages and FTS bug-fixes)
update 33 is being prepared with the ability to publish tool versions
gLite Release News
gLite 3.1 update 32 released to production. The big news is CREAM CE (an older version than the one used in PPS pilot), which is the recommended version for ALICE. It will not be matched by default WMS.
gLite 3.0 update 34 released with a new VOMS server host certificate
for more details, follow the links on the agenda page
EGEE Items From ROC Reports
ROC France asked about progress on GGUS:40565. John answered that the incorrect reporting of MAINT status was caused by timing problems due to the gap between when tests get submitted and when they publish their results. SAM is working on fixing this by saving the maintenance status when tests are submitted in the results DB. Tests will stop returning MAINT as a status, but all visualisation tools will be able to display a consistent picture! We're hoping to have this fix implemented in the next fortnight.
GSTAT is now comparing LFC services comparing Information System and GOCDB. The tests are insignificant in terms of availability and COD alarms; the aim is solely to increase accuracy.
New LFC SAM tests
Explained in the agenda. Roberto said that LHCb needed special LFC tests. John replied that if they were generic, the SAM team would provide them, but since they were specific to LHCb, LHCb should develop their own tests.
gLite 3.0 services to be obsoleted
Check the list! No new updates will be released, and you'll be on your own. Please check, because most gLite 3.0 services are in the list, and it's happening next week.
WLCG Items
On the issue of VOs making direct requests to sites to install software, Steve underlined the policy that things must first be discussed in the Ops meeting, and then mentioned in the VO Card. ATLAS agrees to this procedure (they need libblas library). A two-week timeline will be set, and if a site is unhappy, they can raise it in the following week's meeting.
Steve apologized for the jobs that he ran for GridFest, as he should have sent a broadcast first. A CE in Lancaster broke, several hundred thousand jobs ran, and the RTM demo was very successful.
WLCG issues coming from ROC reports
On the subject of France wondering whether there was a procedure to be notified of changes in the LHC DN alarm list, Torsten said that sites don't need to know, but they could consult the Wiki if they so desired. Since the mails are signed by GGUS, the justification for the request had disappeared.
Upcoming WLCG Service Interventions
Please check the URL in the agenda page for the many interventions this week.
ATLAS Service
No report submitted, nothing special to report
ALICE Service
No one present
CMS Service
No one present
LHCb Service
Roberto repeated the report that he'd given to the 3 o'clock meeting.
Problem at CERN getting enough slots to run (fair share to be investigated).
For the LHCb service master at CERN and Tier1s, only Gridka has broadcast information about tomorrow's 1hr downtime.
Need to review the rules for generating gridmap file. All users in pool should be mapped to a single local user account. Steve said that gridmap file is not supported everywhere and is to become a thing of the past. Roberto will write things down and send to John [his input is now attached to original meeting agenda in LHCb report section].
Storage Service
New Castor versions and SRM interfaces (of interest to RAL and Taiwan).
Tim has been testing new ticketing exchange system. Things are going well, and manual steps should be eliminated. This seemed to please Maria, who nevertheless brought up three tickets which were discussed:
GGUS:37059 - has been fixed according to John McGee, and should be closed
GGUS:39303 - no activity, so Maria suggests to close
Alessandro wondered whether OSG handles Tier2 issues within the cloud differently that in Europe (by going directly to BNL, the Tier1). Maria replied that this was entirely up to them.
AOB
Alessandro had a follow-up query from Wednesday's GDB as to whether the direct assignment of Team tickets to Tier2 would be taken out. "Not at all", replied Steve.
Catalin from Fermilab complained that there had been no follow-up to problems with the BDII and "no compatible resources". Steve requested a written reminder of what the issue was.
Maria advertised the VOMS Admin/VOMRS workshop (21st-24th October) and invited all interested parties to attend. It's F2F, but a phone connection can be arranged.
EGEE Web>SA1>WlcgOsgEgeeOpsMeetingMinutes>WlcgOsgEgeeOpsMinutes2008x10x06 (2008-10-10, JohnShade)
EditAttachPDF
Copyright &© by the contributing authors. All material on this collaboration platform is the property of the contributing authors.