=================================== Minutes from DJRA4.7 status meeting =================================== Date: 1 March 2006 12:00 CEST/11:00 GMT Venue: CERN teleconference +41 22 767 7000 Attendance: UEDIN: Kostas Kavoussanakis (KK), Alistair Phipps (AKP), Davy Virdee (DV) CNRS: Benjamin Vial (BV) DANTE: Loukik Kudarimoti (LK) DL: Mark Leese (ML) Apologies UEDIN: Jeremy Nowell (JN) DFN: Robert Stoy (RS) Agenda ====== * Actions from previous meeting * Workplan: https://edms.cern.ch/document/694396/1 * Holidays/Absences * DONM Currently Active Workpackages ============================= WP1.1 Add DT use cases (closing) WP1.4 Internally review document (closing) WP1.5 Update document as per internal review (closing) WP2.2 Internally review document WP4.2a Design, implement, test Mediator caching WP5.1b Indication of data gaps (closing) WP5.1e Elegant handling of empty results WP5.2 Gather and document NOC/GOC feedback WP6.1 Document service design WP6.2 Document NM-WG v2 mapper design WP6.3 Document security requirements of PerfSONAR (closing) WP7.9 Test e2emonit tool configuration mechanism WP7.14 Modification to allow tests to run to one site instead of all WP8.1 Support LCG e2emonit deployment (Ongoing) WP8.2 Update and maintain e2emonit test deployment (Ongoing) WP8.3 Update and maintain PerfSONAR translation layer test deployment (Ongoing) WP8.4 Update and maintain NPM services test deployments (Mediator, Discoverer, DT, Publisher) (Ongoing) Actions ======= Arising ------- [BV] Import traceroute tool into JRA4 CVS [LK] Provide new GEANT discovery file [KK] Investigate with CNRS if e2emonit node and marianne e2emonit configuration site can continue to run, at least until after EGEE review Continuing ---------- [DV] Fix Publisher bugs #14656 (properties whitespace), #14461 (excess data requests after error), #14489 (reestablish JDBC connection) -- fixed #14656, #14489 and #15257 (reattempt Mediator connection). More testing needed. -- #14461 o/g [AKP] Fix diagrams on NM-WG object model design doc [KK] Discuss pcp with ML or other suitable persons (protocol evaluation) -- will try to discuss with ML on the phone [LK] Follow up NM-WGv2 data access with JANET -- asked, but no response. Will try one more time. Completed --------- Deleted ------- [ML] Document which parameters map to which characteristics (for EGEE2 DT developments involving parameter specification). -- not relevant as NM-WGv2 has specific parameters for each tool at the schema level DJRA4.7 ======= General ------- Good attendance meant a productive meeting - thanks to all who joined. Benjamin will be leaving CNRS on 10/3 - thanks for all his contributions to e2emonit and NPM. PCPd: ML says if it is flexible enough, then GridMon could use it. They have a star topology from RAL (tier 1) to tier 2 sites, and a mesh topology within regional groups associated with a tier 2 site. Follow-up teleconference to take place tomorrow. LK will try again to contact Steve Williams @ UKERNA to find out status of PerfSONAR installation and future plans. If they are not willing to use the current RRD-based PerfSONAR MA, they may be able to use a MySQL-backed PerfSONAR MA (flexible enough to allow any database schema to be used) which is currently in the design stages but due at the end of April (too late to be deployed in time for EGEE-I review!). SWITCH has a PerfSONAR MA service ready, but it's Python-based and so using an newer NM-WGv2 sub-version. A Java release of the new sub-version has not yet been made, just a snapshot. The TL will need to be updated when MAs shift to the new sub-version. This may happen after March, but LK thinks he may be able to make changes to TL with his PerfSONAR hat on - he believes little effort will be required as the changes are just in element chaining. KK is concerned about the different versions and would like to see the data providers supporting the version we settle on. ML says that NM-WG's view is that users are expecting too much from the schemas in terms of stability at the moment and it will be some time until they are stabilised. In the meantime we have to weigh up the pros and cons of different version support, and he thinks that if we require data from different providers using different NM-WGv2 sub-versions, we have to weigh up pros and cons of different version support. It should be noted that the sub-versions are - at the moment - quite similar. LK says that data providers want to know what EGEE do with provided data. If we are really using it in EGEE-II then network providers may be interested in supporting the version we mandate (perhaps in parallel with other versions). On the NM-WGv2 schema, ML says that the focus is on specific schemas for specific tools. There does not seem to be a great demand for more generic metric-based schemas, but they would be added in a similar way to the tool-based schemas. http://stout.pc.cis.udel.edu/NMWG/ is the correct location for up-to-date NM-WGv2 schemas. Balticgrid have been in contact with KK about NPM. KK is following up. Laurence Field has picked up the e2emonit deployment on the Tier 0->Tier 1 star, after prompting by KK. We are asking the middleware groups if they would be happy with iperf results with a limit on the maximum bandwidth measured (to limit impact on the network). No-one on the call had any additional insight on whether this was a good idea. WP1 Report on Network Performance Monitoring ============================================ WP1.1 Add DT use cases ---------------------- LK last week: - improved use case on points raised by AKP and after discussions with DANTE operations RS last week: - wrote up two use cases outlined AKP released an updated document and incorporated summaries of the use cases into the final report. This WP item is closing. WP1.4 Internal review of document (final report) --------------------------------- JN/TS last week: - Reviewed final report This WP item is closing. WP1.5 Update document as per internal review (final report) -------------------------------------------- AKP last week: - Updated document The final report has been submitted for review. This WP item is closing. WP2.2 Internal review of document (spec of interfaces) --------------------------------- TS later: - Review document WP4 Mediator ============ WP4.2a Design, Implement, Test Caching -------------------------------------- AKP next week after: - Design, Implement, Test Mediator caching WP5 Diagnostic Tool =================== WP5.1b Indication of data gaps ------------------------------ DV last week: - Evaluated possibility of adding indication of data gaps to graphs without modifying graph libraries + determined not possible - cannot fix This WP item is closing. WP5.1e Elegant handling of empty results ---------------------------------------- DV last week: - implement fix AKP next week: - system test fix WP5.2 Gather and document NOC/GOC feedback ------------------------------------------ ML will attempt to (but cannot promise): - grab GOC people when/if at RAL - talk to NOC: Steve Williams @ UKERNA - write up WP6 PerfSonar TL ================ WP6.1 Document service design ----------------------------- LK last week: - Produced draft of service design doc AKP last week: - Commented on service design doc LK next week: - Address AKP's comments + complete Note: WP6.2 is higher priority WP6.2 Document NM-WG v2 mapper design ------------------------------------- LK last week: - Produced draft of mapper design doc LK next week: - Complete mapper design document This is the highest priority item as we will be using NM-WGv2 more widely in EGEE-II. WP6.3 Document security requirements of PerfSONAR ------------------------------------------------- Not possible within EGEE-I as PerfSONAR JRA5 have not produced security architecture. This WP item is being obsoleted. WP7 e2emonit ============= WP7.9 Test updated e2emonit tool configuration mechanism -------------------------------------------------------- BV last week: - completed testing WP7.14 Modification to allow tests to run to one site instead of all -------------------------------------------------------------------- BV last week: - adding parameter to iperf tool to space out measurements BV next week: - continue adding parameter to iperf tool to space out measurements WP8 Deployment Support ====================== WP8.1 Support LCG e2emonit deployment (Ongoing) ----------------------------------------------- WP8.2 Update and maintain e2emonit test deployment (Ongoing) ------------------------------------------------------------ AKP last week: - installed new e2emonit on test-03 - fixed bug in secondary producer that caused ~23 days of data to be lost WP8.3 Update and maintain PerfSONAR translation layer test deployment (Ongoing) ------------------------------------------------------------------------------- WP8.4 Update and maintain NPM services test deployments (Mediator, Discoverer, DT, Publisher) (Ongoing) ------------------------------------------------------------------------------------------------------- Holidays/Absences ================= Important dates =============== 31 Mar: Last day of EGEE-I Next meeting ============ Date: Wednesday 8 March 2006 11:00 GMT/12:00 CEST Venue: CERN teleconference +41 22 767 7000 AKP 1 Mar 2006