Difference between revisions of "Wider VO issues"
(→TODO list) |
|||
Line 27: | Line 27: | ||
* VOMS | * VOMS | ||
− | |||
- Use of roles | - Use of roles | ||
- Use of SaroNGS (aow obsolete) | - Use of SaroNGS (aow obsolete) | ||
Line 60: | Line 59: | ||
** Observation of people taking up instructions (feedback). Config UI is complicated. Open UI? | ** Observation of people taking up instructions (feedback). Config UI is complicated. Open UI? | ||
UI with GSI ssh route - but then who to support? | UI with GSI ssh route - but then who to support? | ||
− | |||
= Done = | = Done = |
Revision as of 11:00, 2 October 2014
This is part of the core operations team activites
The task involves:
1. Tracking new VO requests across the Grid and specifically those affecting UK resource providers.
2. Maintaining the extracts and details at: http://www.gridpp.ac.uk/wiki/GridPP_approved_VOs.
3. Tracking the status and problems experienced by the VOs and raising issues to the weekly ops-team meeting - Site enablement problems - User issues
4. Providing (or coordinating) help and support to new VOs who require guidance on becoming grid enabled.
TODO list
* Good practice * For individual users of regional VOs * for medium sized VOs * An example data model - Use checksums - Custodial copies of data - Keep multiple copies for load sharing and resilience - Sending jobs to data - Catalogue synchronisation (Use LFC) - Sensible paths on storage
* VOMS - Use of roles - Use of SaroNGS (aow obsolete)
* How to do stuff - Submitting jobs to data
* Job submission
* Ganga - Stuff for small VOs. * Dirac
Comments to Chris at HEPSYSMAN:
- reliability
- easy of use
- future proofing - LHC VOs dictate directions (e.g. move to CREAM, away from LFC)
- WLCG stack is not complete (holes for data management etc.)
- Hidden features
.... https://wiki.egi.eu/wiki/VO_Services/Services_and_Tools_Portfolio (found via Diane Ganga)
- try not to lose portal and console developments in NGS (worked better with gram on CEs)
- free to support SARONGS in the UK
- Had ngs for a long time but little real work submitted. Should NGS sites have UI point at GridPP WMSes?
- Config changes break things inc. certificates. VO specific tests. What is Nagios support like for multi-VOs?
- Manage VO expectations
- Site admins need to be proactive -> raising VO issues to say TB-SUPPORT or proxy GGUS tickets
- What about other solutions like boinc?
- Typical use cases - should VOs define this when joining?
- Repository of images for middleware. UI. Use of CVMFS.
- Observation of people taking up instructions (feedback). Config UI is complicated. Open UI?
UI with GSI ssh route - but then who to support?
Done
* Resilient VOMS servers
This page is a Key Document, and is the responsibility of Christopher Walker. It was last reviewed on 2013-12-16 when it was considered to be 5% complete. It was last judged to be accurate on (never).