1 / 11

LCG-1 Status

LCG-1 Status. Ian Bird PEB – 27 May 2003. Status. Proceed now with available middleware Cannot wait longer, many issues to address Need to integrate monitoring system (for example) Deploy on certification testbed (under way)

vinny
Download Presentation

LCG-1 Status

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. LCG-1 Status Ian Bird PEB – 27 May 2003 Ian.Bird@cern.ch

  2. Status • Proceed now with available middleware • Cannot wait longer, many issues to address • Need to integrate monitoring system (for example) • Deploy on certification testbed (under way) • Test as much as possible – including WP8/LC’s assistance for 2 weeks • We should get at least the major issues • 2 weeks to package & prepare for deployment • Deploy at the beginning of July • This is an extremely compressed timescale – the deployed system will be mostly untested • But, must begin to run as a service as soon as possible – • this is the only way to learn and to improve the service Ian.Bird@cern.ch

  3. LCG-1 Middleware • Strategy is still and will continue to be: • EDG 2.0 components layered over VDT • Taking whatever components are ready for deployment • To include – as soon as feasible: • R-GMA : we know MDS has severe limitations • RLS – begin with local LRC, migrate to distributed LRCs, RLI, and Replica Manager service • RB • VOMS Ian.Bird@cern.ch

  4. Status of middleware • We have a tagged distribution that has been successfully installed on certification testbed today, but configuration issues: • Gris table empty (#1109) – new – we solved • Pool accounts on SE (#884) – known bug • rm component does not work on SE (#1110) – new – but we solved • MyProxy does not start (#1108) – will be fixed in VDT • R-GMA services not running, queries hang, times out (not in bugzilla?) • … • Still cannot run anything today … (could run RLS with static config files) • Contains all the components that we require in principle • R-GMA still has some inconsistencies and not yet ready for deployment Ian.Bird@cern.ch

  5. VDT • VDT (Globus + Condor) • Version 1.1.8-n – to be delivered this week • Same as 1.1.8 with Globus patches: • MDS – from NorduGrid • Job Managers – to avoid need for shared fs (NFS) • Fixes for gass-cache problem • These fixes are on request of LCG • Will include MyProxy • Currently working with VDT 1.1.8 – • VDT 1.1.8-n will be a “slide-in” replacement – we know exactly what is changed Ian.Bird@cern.ch

  6. Information System • We will use MDS until R-GMA is ready to deploy • MDS • Dynamic – changes to RLS, info providers (better solution – must be shown to work this week) • Static configuration files (this works now, but is a management problem) • R-GMA • Will be in distribution in whatever state • Can be tested and compared in production – depending on the status • Could envisage some RB’s pointing to MDS and some to R-GMA Ian.Bird@cern.ch

  7. Data Management • RLS (from EDG) • LRC – one per VO at CERN • Also possible to run at Tier 1’s but without communication • Add RLI when available • But not in the initial deployment – will not be ready • Data management client tools • Until Replica Manager Service available Ian.Bird@cern.ch

  8. Workflow management • EDG RB • Version 2.0 – significant changes from 1.4.x • Still to be tested – scaling – how many CE’s, RB’s required? • Not just related to number of sites, but also to no. WN’s • Here we need to do significant testing – on the production system • However, the service could be used without the RB Ian.Bird@cern.ch

  9. Subsequent to July • R-GMA • 2 stages : initial deployment with single registry, and then multiple registries (very important) • RLI for RLS – allows distributed RLS • Need migration tools to populate remote catalogs • VOMS – authorization management • GFAL – Posix-like I/O (LCG development) • gcc 3.2.x Ian.Bird@cern.ch

  10. Notes • This is still in-line with original documented plan (WG1) • Delays mean very little time for testing and verification – this will happen on the production service • Will attempt to have 2 weeks pre-deployment with WP8/LC’s • During certification and production – we will find many problems • Request that EDG treat bug fixing as a high priority • We will continuously re-assess where to focus effort • E.g. fixing MDS vs deploying R-GMA, etc. Ian.Bird@cern.ch

  11. Notes - 2 • Problem management • LCG will track all bugs (LCG, EDG, VDT, Globus, Nagios, etc) • We will submit bug reports to appropriate channels (EDG, VDT, etc) • We are already doing this • Testbed support • LCG will still help support the EDG integration TB – at the level of 1 person – should be used appropriately, and he is not dedicated to this Ian.Bird@cern.ch

More Related