London Tier2
From GridPPwiki
| Table of contents |
London Sites
| SITE NAME | CE | SE | Contact
|
|---|---|---|---|
| UKI-LT2-Brunel | dgc-grid-40.brunel.ac.uk dgc-grid-44.brunel.ac.uk | dgc-grid-34.brunel.ac.uk | Henry Nebrensky (henry.nebrensky at brunel.ac.uk) |
| UKI-LT2-UCL-HEP | pc90.hep.ucl.ac.uk | pc55.hep.ucl.ac.uk | Gianfranco Sciacca (gs at hep.ucl.ac.uk) |
| UKI-LT2-UCL-CENTRAL | gw-4.ccc.ucl.ac.uk | gw-3.ccc.ucl.ac.uk | William Hay (w.hay at ucl.ac.uk) |
| UKI-LT2-IC-HEP | ce00.hep.ph.ic.ac.uk hep-ce.cx1.hpc.ic.ac.uk | gfe02.hep.ph.ic.ac.uk | Please use mailing list (lcg-site-admin at imperial.ac.uk) |
| UKI-LT2-QMUL | ce01.esc.qmul.ac.uk | se02.esc.qmul.ac.uk | |
| UKI-LT2-RHUL | ce1.pp.rhul.ac.uk ce2.ppgrid1.rhul.ac.uk | se1.pp.rhul.ac.uk se2.ppgrid1.rhul.ac.uk | Simon George (s.george at rhul.ac.uk) |
- If you want to target all London sites you can use the following requirement in your jdl:
Requirements = Member("GRIDPP-LT2", other.GlueHostApplicationSoftwareRunTimeEnvironment); - The Number of running jobs in London shows you the status in the last week. More details about the monitoring can be found down on this page
Contacting Sites
The preferred mechanism for contacting London sites is GGUS (http://www.ggus.org) since it allows to track the issues. Please make sure to set the site name in the subject line.
- If you have a request for London in general please contact the London Tier2 Manager (mailto:d.colling@ic.ac.uk)
- If you want to have a direct contact with the site admins please use the emails in the table above.
- For London Grid Members please use this contact (https://www.gridpp.ac.uk/tier2/london/private/people.html) list.
Suported virtual organisations (vo's)
This is totally out of date.
- We currently support 21 VO in London, for a detailed site/vo matrix please go at the support table (https://gfe03.hep.ph.ic.ac.uk:4175/vosupport.html) which is updated every day.
ltwo
The LTWO vo is ment to be used by members of the London Universities that are part of the London Tier2. An example of use is for teaching purposes. The students can register to the LTWO VO and do their Grid Exercises.
- If you want to be part of the LTWO vo please go here (https://voms.gridpp.ac.uk:8443/voms/ltwo/) with your certificate loaded in your browser
dzero
- Fredric Villeneuve (f.villeneuve at imperial.ac.uk) is in charge of dZero production for the UK. Please contact him if you have any question concerning dZero in London.
- For more information about dzero please go here (http://www.hep.ph.ic.ac.uk/~villeneu/samgridlcg/samgridlcg.html)
Development and Testing
SGE Integration
Imperial has produced an implementation of the JobManager and Information plugin for SGE. We currently have three ce using it in production. Keith Sephton (kms at doc.ic.ac.uk) has taken over the work of David McBride who did the original implementation. Details about the implementation can be found at LCG-on-SGE
Monitoring
We have developed a monitoring tool called Grid Load (http://gridportal-ws01.hep.ph.ic.ac.uk/gridload/00_example_webpage.html) that allows to see the number of jobs in any given state. For more details please see the Readme (http://gridportal-ws01.hep.ph.ic.ac.uk/gridload/00_PLEASE_README.html).
GLite Pre Production Service (PPS)
Barry McEvoy (http://www.hep.ph.ic.ac.uk/e-science/people/macevoy.html) is installing a Glite PPS. He has currenlty installed a CE,UI,MON,WMS,5 WN
Experiment Involvment
SC4
More information about SC4 can be found here. The current plans and activities in London Regarding Service Challenge 4 can be found at London SC4 Activity
CMS
- CMS Activities in London are lead by D.Colling. We are in the process of having all London sites active for the CMS montecarlo production. One PheDex (http://twiki.cern.ch/twiki/bin/view/CMS/PhEDEx) agent is running for the whole London Tier2 (gfe03.hep.ph.ic.ac.uk).
- For More information about the CMS activities in London please go to LT2_CMS
Other Activities
Cross Site Support
The objective of cross site support is to improve our coverage in case a site-admin is not there and a site admin from another site could help. More detail on the Cross site support can be found at this (https://www.gridpp.ac.uk/tier2/london/public_keys/index.html) location
Blog
A blog is kept with what is happening in London. This is the place to look for day-day news. The blog is at http://londongrid.blogspot.com/. We also give an rss feed (http://londongrid.blogspot.com/feeds/posts/default) for the blog.
BandWidth Tests
The bandwidth Test have been achieved using Network_Testing
RB data analysis
The Resource Broker (RB) data analysis is a set of root (http://root.cern.ch) macros designed to analyze the ressource broker data produced by the Real Time monitor (http://gridportal.hep.ph.ic.ac.uk/rtm/).
Security
- Site security contacts can be found in the goc (https://goc.grid-support.ac.uk/gridsite/gocdb2/index.php) database
- Grid Security Policy document: [1] (https://edms.cern.ch/document/428008/4)
- Security Feed: http://rss-grid-security.cern.ch/rss.xml
Configuration Tips
Calendar
- Upgrade schedules/downtime plans: google calendar (http://www.google.com/calendar/embed?src=kovf3259830mm69dioh08e7ejc%40group.calendar.google.com&pvttk=97a28aa948be35aa2805e8fd2190fdaf)
Resources files
- Summary of the London resources: XLS (https://www.gridpp.ac.uk/tier2/london/Data/lt2-ressources-summary.xls)
- Full xls sheet of the resources usage: XLS (https://www.gridpp.ac.uk/tier2/london/Data/LT2-Dashboard.xls)
VO configuration
- The following LondonGrid-yaim-vo can be appended to the yaim configuration file in order to support the VOs recommended by GridPP. Beware that it assumes DPM_HOST to be your SE. The queue part also needs to be properly configured for your site. The contents of the vo.d directory is LondonGrid-vo.d.
- The LondonGrid-group.conf should also be correctly populated
General Links
- Freedom of Choice Excluded (FCR) link (http://www.hep.ph.ic.ac.uk/~aggarwa/fcr/fcr.html). Python script for UK sites here.
- London Tier 2 GriddPP home page (http://www.gridpp.ac.uk/tier2/london) and the blog (http://londongrid.blogspot.com)
- How to send a mail to a given DN see the CIC portal (https://cic.gridops.org/index.php?section=roc&page=usertracking)
- How to broadcast (https://cic.in2p3.fr/index.php?id=rc&subid=rc_publish&js_status=2) a message
- How to export your ssl keys here (http://www.grid-support.ac.uk/content/view/34/35/)
- Relocatable Glite tarball (http://grid-deployment.web.cern.ch/grid-deployment/download/relocatable/) for WN and UI
- Certificate Authority RPMS (http://grid-deployment.web.cern.ch/grid-deployment/lcg2CAlist.html)
- FootPrint ticketing system: http://helpdesk.grid-support.ac.uk/
- Security Practices (https://cic.in2p3.fr/index.php?section=roc&page=securityissues)
Meeting Links
The meeting Agenda is kept here (http://agenda.cern.ch/displayLevel.php?fid=338). Direct links to previous meetings can be found below.
- 25/01/2006: Technical Meeting (http://agenda.cern.ch/fullAgenda.php?ida=a06661)
- 22/03/2006: Management Board Meeting (http://agenda.cern.ch/fullAgenda.php?ida=a061785)
- 22/03/2006: Technical Meeting (http://agenda.cern.ch/fullAgenda.php?ida=a061786)
- 31/05/2006: Technical Meeting (http://agenda.cern.ch/fullAgenda.php?ida=a062587)
- 05/07/2006: Technical Meeting (http://agenda.cern.ch/fullAgenda.php?ida=a063159)
- 09/08/2006: Technical Meeting (http://agenda.cern.ch/fullAgenda.php?ida=a063262)
- 13/09/2006: Technical Meeting (http://agenda.cern.ch/fullAgenda.php?ida=a063280)
- 25/10/2006: Technical Meeting (http://agenda.cern.ch/fullAgenda.php?ida=a063465)
- 15/11/2006: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=8673)
- 12/12/2006: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=9691)
- 17/01/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=11163)
- 09/02/2006: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=9691)
- 12/02/2007: Discussion about the review. No indico agenda, minutes in the lt2 archives.
- 28/03/2007: Management Board Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=14434)
- London Tier2 Review 17/04 -- 20/04
- 02/05/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=15539)
- 13/07/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=17965)
- 23/07/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=19549)
- 30/07/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=19202)
- 06/08/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=19545)
- 13/08/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=20039)
- 20/08/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=20184)
- 17/09/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=21394)
- 29/10/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=23531)
- 05/11/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=23530)
- 26/11/2007: Technical Meeting (http://indico.cern.ch/conferenceDisplay.py?confId=24650)
Actions
(Olivier): Update the contact list.060705-1 (Olivier) 13/09/2006: Circulate sudo files to enable cross support at sites that are happy with that.- Solution: Example of sudoers file here (https://www.gridpp.ac.uk/tier2/london/public_keys/sudoers)
060705-2 (Giuseppe) 13/09/2006: Find out if Alex is happy to give sudo access to se, ce and mon.- Solution: Alex Agreed to give sudo for the ce/se/mon
060809-1 (Duncan) 06/08/18:: Find out why the crl tests are sometimes failing and what it really does- Solution: I think these crl !!! errors occur when you submit a job manually from the sft submit page (https://monitoring.egee.man.poznan.pl/admin2/). GGUS ticket has been submitted to ask to solve that problem. ref: GGUS 11240
060809-2 (Duncan) 06/08/18:Find out what is the conclusion about lhcb submitting jobs as lhcbsgm. It it a configuration problem on our side our on their side.- Solution: Ricardo Graciani states he needs access from some jobs as lhcbsgm for debugging and SW installation; should be sorted when VOMS is fully operational. For more details see end of GGUS ticket 11307 (https://gus.fzk.de/pages/ticket_details.php?ticket=11307&from=fct)
060809-3 (Olivier) 06/08/17:Make a wiki entry for SC4 activities in London- Solution: Done can be found at London_SC4_Activity
060809-3 (Olivier) 06/08/17:Ask Matt Hodges what problems have been seen on their raid arrays to help QMUL.- Solution: Giuseppe said that putting new disks solved the problem.
- 060809-4 (Olivier): Sort out the problems with installing apt-get on lesc. Lesc is using 64bit rpm database
060809-5 (Giuseppe) 06/08/17:Solve the Apel accounting problem at QMUL. Will have to install the rpms provided by Dave Kant on the CE and MON.060809-6 (Keith): 06/08/25Dzero has submitted 200 Jobs and I (ovda) have never seen more than 40 running. Check that everything is correctly configured in sge. Kept open to understand why some of the s072 queues are marked as au- Solution: monitoring consistent with what is seen in the cluster. The queue conf had been modified to accept more jobs for Grid.
- 060809-7 (David): Find out what are the plans for Atlas with S. Lloyd to proceed in LT2
060809-8 (David):Get in touch with Brunel to help in sorting out the 1.5Mb/s bandwidth problem. Ongoing060814-1 (Mona):Check that the queues on the SGE cluster are set to closed when the site is in downtime- Solution: We can only set the whole site down. Not on a per queue basis.
060913-1 (Olivier):Check the Biomed voms settings- Solution: Checked with Yannick Legre and agreed with the voms settings. Now shown in the LT2 wiki.
060913-2 (William):Compare the figures of the GridLoad Plots (https://gfe03.hep.ph.ic.ac.uk:4175/cgi-bin/load) with the number of running jobs at UCL- Solution: Done and is ok.
060913-2 (Duncan):Ask Gstat to change the scale of their plots to be logarithmic- 061025-1 (All):Comment the transfer test page (http://www.gridpp.ac.uk/wiki/Service_Challenge_Transfer_Test_Summary)
- 061025-2 (David, Austin, Keith, William, Giuseppe):Add your public keys at https://www.gridpp.ac.uk/tier2/london/public_keys/
- 061025-3 (Austin):Finalize the read/write transfer tests for UCL (HEP, CENTRAL)
- 061210-1 (Olivier): Ask Graeme for the routing table on the worker nodes to avoid going trough the nat.
061212-1 (Giuseppe):Cross site support: Ask Alex if using ltwosgm is ok.- 061212-2 (Gianfranco):Cross site support: Ask Ben/Gordon if ok to create accounts for DC,OV,MA,GM,DR,GS,WH
- 061212-3 (Keith): Cross site support: Circulate to point out the link to register
061212-4 (Olivier):Cross site support: Give list of commands that sudo can run for the CE,SE.- 061212-5 (Olivier): Cross site support: Create a wiki entry for the cross site support.
- 061212-6 (Olivier): See how to publish main daemon state on a web page.
- 061212-7 (All): Update the network diagram on the 10 Easy nework question (http://www.gridpp.ac.uk/wiki/GridPP_Answers_to_10_Easy_Network_Questions) page.
- 0701??-1 (David): Check on upgrade plans for London MAN.
