Core-ops tasks

Thursday 31 Jul 2014, 11:00 → 13:30 Europe/London

EVO - GridPP Operations team meeting

- This a meeting for the review of the GridPP ops core tasks - The intention is to run the meeting in Vidyo. Direct link http://vidyoportal.cern.ch/flex.html?roomdirect.html&key=HRHLwpRDyPg7. Pin 1234. - To join via phone see http://information-technology.web.cern.ch/services/fe/howto/users-join-vidyo-meeting-phone for dial in numbers. -- The London (UK) service is on +442030510622 -- The meeting extension is 9313701. Apologies: Matt D.

notes July2014-core-ops-notes.pdf

11:00 → 11:10 Documentation

https://www.gridpp.ac.uk/wiki/Documentation Done in Q2: - TBC To discuss: - https://www.gridpp.ac.uk/php/KeyDocs.php -- Various changes made but none updating. - Review EGI OLA/SLA updates: https://indico.egi.eu/indico/materialDisplay.py?contribId=2&materialId=slides&confId=2191 Blogs: http://planet.gridpp.ac.uk - Can we break home page aggregation into GridPP and non-GridPP to stop our posts disappearing so quickly? Notes from last meeting: - keydocs may need a rethink. Something more active & organic - Planet (apache server) works but header issue. Will rejig ScotGrid website... see if something more useful than planet inc. RSS and Tumblr (A. Talk to Tom)

11:10 → 11:20 Monitoring

https://www.gridpp.ac.uk/wiki/Monitoring Done in Q2: - Collecting UK feedback and representing in consolidation group To discuss: - who is using site nagios? - sharing scripts... is there more to be done? - where does puppet/config fit - ECDF-Glasgow joint monitoring? Notes from the last meeting: - Nagios plugins. Testing at site ongoing. Graphite in use. Graphene dashboard uses Graphite. (interest from Pete, Robin & Chris). - Wiki page links to documentation - Github focus (where does this fit best? Under SR or tools?): - For links page some updates needed - Bristol/Oxford/Robert sharing. Meeting at HEPSYSMAN suggested. - Need ARC CE modules. Post to blog on process. - Suggest monitoring session at HEPSYSMAN - advance puppet usage.

11:20 → 11:30 Staged rollout

https://www.gridpp.ac.uk/wiki/Staged_rollout Done in Q2: - EMI-3 completed(ish) -- new reports in https://www.gridpp.ac.uk/wiki/Staged_rollout_emi3? - involvement in middleware readiness work. Notes from last meeting: - not many updates - Andrew L tested ARC CE

11:30 → 11:40 Core services

https://www.gridpp.ac.uk/wiki/Core_Grid_services Done in Q2 2014: - perfSONAR - upgrades to installs. - IPv6 investigations - VOMS - https://voms.gridpp.ac.uk:8443/vomses/ -- upgraded? Testing -- IPv6 (Glasgow, Oxford, IC...) - continued progress Notes from last meeting: - Discussion about reliability of PS boxes

11:40 → 11:50 Wider VOs

https://www.gridpp.ac.uk/wiki/Wider_VO_issues Done Q2 14: - Expanding CERN@school activities - Further enabling and supporting ABC - Testing of DIRAC (but less so than desired) What next? - Push wider WebDAV usage? https://www.gridpp.ac.uk/wiki/WebDAV#Federated_storage_support - Still ... 'quick start' documentation

11:50 → 12:00 Regional tools

Done in Q2 14: - Presentation/testing with DIRAC What next? - Still push DiRAC testing! - Make VO Nagios more useful? Notes from last meeting: - Wider use of VO Nagios (A: KM to write blog article on topic) - Changes to central SAM services - DIRAC - need to be clear on state of play in EGI and how this relates to UK - Should we 'jump', or get involved with central DIRAC instance?

12:00 → 12:10 Interoperation

https://www.gridpp.ac.uk/wiki/Grid_interoperation Done Q2 14: - Continued representation at EGI ops meetings What next? - (DC cloud/technical discussions and ops overlaps)

12:10 → 12:15 Security

https://www.gridpp.ac.uk/wiki/Security Done Q2 14: - Following up on various security issues - Security training at HEPSYSMAN - ARGUS cascade - CA moved to SHA-2 Issues/concerns - Some areas not getting attention such as reviewing cloud approaches - Security dashboard lost Notes from last meeting: - Glue2 - confusion about what is working vs in error. Problems transient. - Include in ticket a link to why/when error was noted - common errors in glue2 are default values leading to timeout - (A. EGI)

12:15 → 12:20 Accounting

https://www.gridpp.ac.uk/wiki/Accounting Done Q2 14: - Support for EMI-3 APEL fixes... What next? - Broaden area to 'New technologies and impacts'? (e.g. use of whole node or cloud scheduling, impacts of many core....) -> testing? - Other metric areas - Get involved with APEL testing Notes from last meeting: - not as many VMs as previously (A. mail Peter) - authentication issues - Resource reporting: VAC resources reported; in Panada monitoring & ATLAS ganglia - Accounting: GridPP accounting uses Panda. Other expts. need to make sure in APEL. - Cloud queues: VAC one per allocation. (A. check cloud accounting for Imperial) - REBUS: not publishing for OpenStack

12:20 → 12:22 Ticket follow-up

https://www.gridpp.ac.uk/wiki/Ticket_follow-up Done in Q2: - Steady review of tickets and chasing where needed. - Outstanding ticket numbers decreasing

12:22 → 12:30 The next month for operations...

Notes from last meeting: - multi-core: CMS not running yet. ATLAS no stable stream. - QMUL happy with SGE implementation. RAL with HT Condor - Discussion on propriety software. - Focus for next month - ARGUS - WN tarball - Multi-core (incl. monitoring) - Other things to follow-up: - ATLAS RIPE anchor - DIRAC testing/usage

12:30 → 12:31 AOB

- Tentative next meeting date Thursday 28th August (i.e. after GridPP33). - GridPP4+ - EGI is updating RC OLA (comments by 15th Aug): https://documents.egi.eu/public/ShowDocument?docid=31.