WLCG-OSG-EGEE Operations Minutes Mon 10 Dec 2007

• Date: Mon 10 Dec 2007
• Agenda: 23802
• Chair: SteveTraylen
• Minutes: NicholasThackray
• Minutes from previous meetings: WlcgOsgEgeeOpsMeetingMinutes
• NOTE: Extracts from the agenda are shown in italics.

Attendance

EGEE

• Asia Pacific ROC: Absent
• Central Europe ROC: Marcin
• OCC / CERN ROC: John Shade, Antonio Retico, Nick Thackray, Steve Traylen
• French ROC: Gilles, Piere-Emanuel, Rolf, Helene
• German/Swiss ROC: Clemens Koerdt, Sven Hermann
• Italian ROC: Alessandro, Paolo
• Northern Europe ROC: Absent
• Russian ROC: Lev
• South East Europe ROC: Kostas
• South West Europe ROC: Absent
• UK/Ireland ROC: Jeremy, Andy Newton
• GGUS: Thorsten
• OSCT: Absent

OSG

• OSG GOC: Rob

WLCG

• WLCG Service Cordination: Absent
• Grid DB services: Absent

WLCG Tier 1 Sites

• ASGC: Absent
• BNL: Absent
• CERN site: Ulrich
• FNAL: Jo Kaiser
• FZK: Clemens
• IN2P3: Piere
• INFN: Alesandro, Alfrede
• NDGF: Absent
• PIC: Absent
• RAL: Derek
• SARA/NIKHEF: Absent
• TRIUMF: Rod Walker

Reports Not Received

• VOs: ALICE, ATLAS, CMS, LHCb
• EGEE ROCs (Prod Sites): AsiaPacific, France, Russia, SWEurope
• EGEE ROCs (PPS Sites): AP, FR, IT, NE, RU, SEE

Feedback on Last Week's Minutes

None were given.

EGEE Items

Grid Operator Hand Over on Duty

	Primary Team	Secondary Team
From	ROC South West Europe	ROC UK/Ireland
To	ROC France	ROC Central Europe

• Nothing to report.

PPS Reports

Issues from EGEE ROCs:

SAM UI at Cyfronet has been configured to submit jobs also to uncertified PPS sites(ticket #28436). There was a problem with permissions on lfc02.pic.es, which cause failure of RM tests for all sites - fixed (ticket #30035). [CE ROC]

No further questions.

Release News:

gLite 3.1.0 PPS Udate 11 is about to be released to PPS sites (due today). Notably it contains, together with the usual bug fixes, the new VOBOX service for SL4 (32 bit) glite-yaim-core has been released, all the metapackages have been updated.

No further questions.

EGEE Items From ROC Reports

ROC-DECH: SAM tests for SRM are currently only submitted every two hours. Can this frequency be doubled?

OCC will follow up for next meeting. [Action 71]

Q: What is the reason for increasing the frequency?
A: To reduce downtime of SRM (which is not so stable) when there are any problems.

gLite Release News

gLite 3.1 Update 07 included patch #1389, an update to GFAL/lcg_util. A serious problem has been found with this patch, whereby lcg-cr segfaults with a classic SE endpoint: GGUS ticket 32016. Consequently, this patch has been removed from the production repository. Sites which have not yet upgraded will not be affected, but sites which have already upgraded to the affected rpms should do the following: # rpm -e --nodeps GFAL-client lcg_util CGSI_gSOAP_2.7 # yum update glite-WN This will roll back to the earlier versions. For the record, the rpms removed are: * GFAL-client-1.10.5-1.slc4.i386.rpm * lcg_util-1.6.4-1.slc4.i386.rpm * CGSI_gSOAP_2.7-1.2.1-2.i386.rpm The release team apologises for this situation.

It should be noted in the minutes that this problem was not seen in the PPS as this update bypassed the PPS (as requested by LHCb and agreed at the EMT).

After some discussion it was decided that the production release notes should clearly indicate when an update has bypassed the PPS. [Action 72]

Proposal to stop using dteam VO in SAM monitoring

For historical reasons some of the standard, regular submissions of SAM tests are being carried out under the dteam VO. We would like to stop doing this and have everything under the OPS VO by next Monday (17 December). Anyone with an objection to this should contact sam-support@cernNOSPAMPLEASE.ch before Friday 14 December.*

No further questions at the meeting.

Upgrade of CERN AFS UI from gLite 3.0 to gLite 3.1

As announced on Monday, 22 Oct 2007, the default version of the AFS UI will be changed from gLite 3.0 to gLite 3.1. In practical terms this means that the 'current' AFS UI link will not point to the latest 3.0 version anymore, but to the latest 3.1 version. The change will happen this Wednesday, 12 Dec, 2007 10:00 CET (09:00 UTC) An EGEE broadcast will also be sent out to announce the change.

There were no further questions.

WLCG Items

WLCG issues coming from ROC reports

None.

Upcoming WLCG Service Interventions

1. LFC at CNAF will be upgraded to 1.6.7 and Oracle streams from CERN reconfigured to replicate also CNS_userinfo and CNS_groupinfo tables there. This will take place on Thursday and Friday, 13 & 14 December.

FTS Service Review

See the report attached to the agenda. In particular INFN, RAL and PIC should read the report.

ATLAS Service (Alessandro)

Report at the meeting: ATLAS had some sites in Milan that removed a service, but it continued to be reflected in availability & reliability calculations. Steve explained that the service should be put into maintenance or downtime beforehand, so that the last SAM test will have the correct status before the service is removed. Alessandro suggested that the info be re-circulated.

ALICE Service (No representation)

No report.

CMS Service (No representation)

No report.

LHCb Service (Roberto)

Roberto reported that the LFC intervention at CERN was successful. He also reiterated the scheduled intervention at CNAF for LFC.

WLCG Service Coordination

Nothing this week.

OSG Items

GGUS ticket 28485 is outstanding. Rob says that they are awaiting a reply from ATLAS. He will chase this up again.

Review of Action Items

Click HERE for the list of actions.

AOB

• Intervention on the LFC service at CERN proceeded without incident.
• Intervention on the VOMS/VOMRS service at CERN proceeded without incident.
• There will be a modification to the service Core flag in the GOC database. This is to allow the further refinement of the target audience for automated broadcasts. Unless anyone disagrees, the change will take place on Wednesday, 13th December.
  Two weeks ago, some people complained about receiving notifications about downtimes not relevant for them. This was because when a downtime impacts a core node our notification system, by default, informs all ROC, RC and VO managers. Therefore, in agreement with the ROC managers, we proposed a solution whereby the GOC DB Core flag of a service could take the values GRID, FEDERATION AND VO (previously it was a Y/N flag).
  So, if the Core flag of a service is correctly set, when a downtime impacts:
  • a VO's node our system will contact:
    • all VO supported on this node
    • Site Administrator at which belongs this node
    • ROC managers at which belongs this node
  • a FEDERATION's node our system will contact :
    • all VO supported on this node
    • all Site Administrators of the impacted region
    • ROC managers at which belongs this node
  • a GRID's node our system will contact :
    • all VO supported on this node
    • all Site Administrators of the grid
    • all ROC managers of the grid

It is very important that all sites correctly set the status of their "core" services with respect to what type of core service they are. Only when this is done will the new system work effectively. By default all nodes which had the Core flag set to YES are now set to GRID.

Next Meeting

The next meeting will be Monday, 17 December 2007 15:00 UTC (16:00 Swiss local time).

• Attendees can join from 14:45 UTC (15:45 Swiss local time) onwards.
• The meeting will start promptly at 15:00 UTC.
• The WLCG section will start at the fixed time of 16:30.
• To dial in to the conference:
  • Dial +41227676000
  • Enter access code 0157610

---

These minutes can only be changed by members of:

• Set ALLOWTOPICCHANGE = WLCGOpsMeetGroup