Upload
chaeli
View
30
Download
0
Tags:
Embed Size (px)
DESCRIPTION
HEPSYSMAN May 2007. Physics IT Services PP Computing. Oxford & SouthGrid Computing Status (Ian McArthur), Pete Gronbech May 2007. Physics IT (windows services). - PowerPoint PPT Presentation
Citation preview
HEPSYSMANHEPSYSMANMay 2007May 2007
Oxford & SouthGrid Oxford & SouthGrid
Computing StatusComputing Status(Ian McArthur), Pete Gronbech May 2007(Ian McArthur), Pete Gronbech May 2007
• Physics IT Services
• PP Computing
Physics IT (windows services)Physics IT (windows services)
Vista Deployment: We intend to deploy Vista after Service Pack 1 is released. Most Desktop Systems will need to be upgraded to take advantage of the new features. ( More RAM / Graphics Card Replacement ).
Office 2007 Deployment: Office 2007 will be deployed as soon as time allows.
Internet Explorer 7: We are delaying the release of IE 7 to managed desktops until we receive confirmation from the OSRIS ( Financial ) support team that IE7 is supported. The deployment may be brought forward if the number of 0-day vulnerabilities increase though.
Spysweeper: Spysweeper has now been removed from Managed desktops and laptops. We are now relying on Sophos exclusively for Anti-Virus and Malware detection and removal. If you find that you PC is running slowly then please email NT Support so that we can investigate why.
Diskserver Replacement: The user areas on PPFS2 will be moved within the next few months to new disk storage. Disk quotas will be increased.
Campus Common DesktopCampus Common Desktop
We have been evaluating the use of Altiris to allow deployment of virtual applications. This simplifies rollout and updating of software by rolling the applications whole environment into a single package which can be downloaded and run by any user. We may be able to share the generation of packages with the Common Desktop support staff.
Email Anti-SpamEmail Anti-SpamQuarantined Emails
0
500
1000
1500
2000
2500
3000
3500
4000
1 3 5 7 9 11 13 15 17 19 21 23 25 27
Day number (Sep 06)
Em
ail
s p
er
da
y
Quarantine Release Requests
0
10
20
30
40
50
60
70
80
1 3 5 7 9 11 13 15 17 19 21 23
Day (Sep 06)
Re
lea
se
Re
qu
es
ts p
er
da
y
Quarantined Emails
0
500
1000
1500
2000
2500
3000
3500
4000
1 3 5 7 9 11 13 15 17 19 21 23 25 27
Day number (Last 28 days)
Em
ail
s p
er
day
Quarantine Release Requests
0
10
20
30
40
50
60
70
80
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Day (Last 28 days)
Re
lea
se
Re
qu
es
ts p
er
da
y
Sep 06
Now
Incoming SPAM rate has increased slightly but quarantine release rate has fallen a little. Very few released emails actually turn out to be useful emails.
SharePoint
An installation of MS SharePoint Services is being evaluated, in particular by the EURECA project (Hans Kraus)
•Web browser-based team-working software•Provides workspace for:
•Schedules•Documents•Discussions•Wikis, blogs, etc.
•Can be administered by members of research group•Available to users outside Physics
Computer Rooms - BegbrokeComputer Rooms - Begbroke
Department has signed up to share the new computer room at Begbroke Science Park with the Oxford Supercomputer.
We will get 22 out of the 55 rack spaces and can generate up to 250KW of heat.
This will provide sufficient space for the PP tier-2 centre, new large clusters for Astro and Theory plus a number of smaller clusters (AOPP and Clarendon)
The Computer room is housed in the Institute of Advanced Technologies (AIT) which will be opened by Alistair Darling MP on 15th May. Work on the computer room starts immediately afterwards and should be complete by mid-August.
There will be a couple of weeks of acceptance testing before handover and installation during September.
Currently working through the detailed design with the chosen main contractor before finalising plans and costs by early May.
New Begbroke Computer RoomNew Begbroke Computer Room
The New Computer room being built at Begbroke Science Park jointly for the Oxford Super Computer and the Physics department, will provide space for 55 (11KW) computer racks. 22 of which will be for Physics. Up to a third of these can be used for the Tier 2 centre.
Computer Room – Local InfrastructureComputer Room – Local Infrastructure
Even with space at Begbroke, we need space to hold our infrastructure and development machines.
Existing computer rooms are poorly equipped with both power and air-conditioning and often take up valuable office space.
We are therefore building a new computer room on level 1 of DWB.
It will hold up to 20 racks generating up to 100KW of heat. (There maybe an option to install more air conditioning if needed). This is a small fraction of the load that would have been generated for a shared physics/OSC room in the basement (700KW+).
Room should be ready by June and in the short term can house some equipment destined for Begbroke (e.g. the PP tier-2).
Local DWB Computer roomLocal DWB Computer room
Local Physics department Infrastructure computer room (100KW) has been agreed. This will relieve local computer rooms and possibly house T2 equipment until the Begbroke room is ready. Racks that are currently in unsuitable locations can be re housed.
Particle Physics ComputingParticle Physics Computing
Staff Changes– Ewan MacMahon – New Unix System Administrator for PP– Sigal Cohen - New DAQ software developer
PP linux Desktops – Ewan will be evaluating SL4, and possibly other distributions for suitability as a
desktop Linux– It is essential that it is easy to maintain, and update in a secure way
PP Cluster– New Server purchased for Linux file serving (pplxfs2)– Data was migrated in Dec 06– New infrastructure/installation/monitoring server– Userdisk upgraded to RAID array of 6 146GB disks (April 07)
Local Grid Resources. Tier-2 Cluster– No new purchases due to lack of computer room / Electrical Power
PP ClusterPP Cluster
The new server (dual-core, dual socket AMD Athlon 285 with 8GB ram) to replace pplxgen called pplxgenng is running Scientific Linux 4. We are planning to migrate all the servers to SL4 but need to wait for CERN etc. Users can test their applications on this machine now.
Cluster load during April 07 by group
• 31 CPUs for general interactive work, development, testing and small scale batch runs.
Group % of load
Lhcb 25.5
Sno 24.7
Atlas 16.26
Harp 12.39
Lc 12.17
Licas 3.51
Zeus 3.4
Minos 1.29
PP cluster load Q1 07
PP Tier-2 PP Tier-2 Currently 40 dual processors. Expansion delayed due to shortage
of computer room facilities. Have plans and funding to purchase ~50 dual-core, dual socket
servers (comparable to 200 CPUs). Oxford supports 20 VOs. 17 of which have run jobs in the last year. Most active VOs are LHCb (38.5%), Atlas (21.3%) and Biomed
(21%). 300,000 kSI2k hours delivered in the last 12 months.
CMS, Atlas and LHCb have been the heavy users during April
The graph on the right shows usage for the last year
Status at CambridgeStatus at Cambridge
Currently glite 3 on SL3 CPUs: 32 2.8GHz Xeon 3 TB Storage
– DPM enabled Oct 05
Upgrade arrived Christmas 2006 32 Intel ‘ Woodcrest’ based servers, giving 128 cpu cores equiv. to approx 358 KSI2k.
Local computer room upgraded.
Storage upgrade to 40-60TB expected this summer.
Condor version 6.8.4 is being used but the latest LCG updates have a dependency for condor-6.7.10-1. This development release should not be used in a production environment. LCG/glite should not be requiring this release.
Cambridge (2)Cambridge (2)
CAMONT VO supported at Cambridge, Oxford and Birmingham. Job submission by Karl Harrison and David Sinclair
LHCb on Windows project (Ying Ying Li)– Code ported to windows
• HEP 4 node cluster• MS Research Lab 4 node cluster (Windows compute cluster)
– Code running on a server at Oxford, possibly expansion on OERC windows cluster
– Possible Bristol nodes soon
Other Southgrid sitesOther Southgrid sites
RAL PPD , Birmingham and Bristol own talks
Other groups within the Southgrid EGEE area are; EFDA-JET with 40 cpus up and running
The Advanced Computing and Emerging Technologies (ACET) Centre, School of Systems Engineering, University of Reading started setting up their cluster in Dec 06.
Steve Lloyd TestsSteve Lloyd Tests21.3.0721.3.07
Site MonitoringSite Monitoring
Grid wide provided monitoring– GSTAT– SAM– GOC Accounting– Steve Lloyds Atlas test page
Local Site Monitoring– ganglia– pakiti– torque/maui monitoring CLIs– Investigating MonAMI
Developing– Nagios; RAL PPD have developed many plugins, Other
SouthGrid sites are just setting up
SummarySummary
SouthGrid continues to run well, and its resources are set to expand throughout this year.
Birmingham new University Cluster will be ready in the Summer.
Bristol small cluster is stable, new University cluster is starting to come on line.
Cambridge cluster upgraded as part of the CamGrid SRIF3 bid.
Oxford will be able to expand resources this Summer when the new computer room is built.
RAL PPD has expanded last year and this year, way above what was originally promised in the MoU.