COD DECH Meeting

Friday 22 Jun 2007, 11:00 → 12:00 Europe/Zurich

Phone conference

Clemens Koerdt

Regional Meeting of the COD DECH Team members

minutes COD DECH 22.6.2007 attendance victor, christian, clemens, alessandro COD-13 Stockholm ============== Collection of point of interest from the meeting: - SAM portal team going to migrate to GridView (timeline 3 months) - advantages: more suitable approach to group historical results and grouping of results per site - CIC dashboard is going to implement GridView bars on the same timeline - SAM already has a web interface to modify the automatic masking of alarms: http://lcg-sam.cern.ch:8080/alarms/mask_alarm.xsql - A similar web interface will be made available for weighting of alarms - People at the meeting agreed that a review of the prioritisation mechanism for the alarm is necessary. - your opinion is welcome. - CIC dashboard has a new feature: now possible to send emails to sites without escalating ticket at the same time. Christian: It would be nice to be able to set a new expiration date for the ticket in the same step. - a new downtime procedure was proposed by Osman, which combines declaration and broadcasting in one step - Failover of the critical grid portals is progressing - CIC already in place at CNAF, some improvements ongoing - SAM failover at Cyfronet to be tested soon. - GOCDB3: web frontend in place at ITWM, work ongoing in replicating the database as well - GGUS failover mechanism has also developed and another failover at University will be tackled later - Work of the TIC (Tool Improvements for COD) working group : - new proposal for CA update procedure presented and appreciated by COD people - a couple of improvements in test documentation and increased verbosity on possible central failures in the error messages (esp. rm-tests) - Web site: http://goc.grid.sinica.edu.tw/gocwiki/Tools_Improvements_for_COD - New Working Group to be let by COD DECH: Best practices (see slides on the COD-13 agenda page for more details) - aim is to collect recommendations to improve the work (and credibility of the COD teams) - to provide input for the Operations Manual - there have been some open issues from the Operations Manual working team: - handling of tickets to PPS-sites: proposal was to stop issuing tickets to PPS sites under current conditions - found a couple of use cases to open ticket from the dashboard, which are not necessarily related to an alarm (can be found under 'Other Tools' on the dashboard. - adopted proposal for a revised procedure to follow when introducing new sam tests - pointed out the large number of steps involved to open a single ticket (some recommendations to CIC team) - ... Internal Organization =============== - checked availability of people for the coming shifts - some people are looking for replacements during summer months (will be followed up via email on our list) - informal discussion on possible reorganization of shifts: sharing the work not by time and/or site, but rather by concentration on specific tests - advantages may be a more effective work, increase expertise due to specialisation, might also help in detecting trends, .... - active every shift instead of every second shift, but with reduced intensity Clemens and Victor expressed a certain preference for such a kind of organization Alessandro and Christian expressed their reservation to the proposal: - to much specialization is bought with reduced general overview, which might be more important - work sharing as it is done today works rather well - lower intensity, but longer periods of work not well appreciated - it was not intended to take a decision on such a reorganization today. - discussion will be picked up again in a forthcoming meeting. - people were encouraged to forward any issues they see with tools, tests or workflow, etc. to the mailing lists and/or the appropriate COD working group (SAM, CIC, TIC, ...) - people were also encouraged to get involved in writing GOC wiki entries and particularly http://goc.grid.sinica.edu.tw/gocwiki/SiteProblemsFollowUpFaq - the latter triggered some discussions on the need of documentation from the developers/release teams as well as improved error messages and log files and the like. Related to this, please find here the link to the latest SA1 workshop in Stockholm, where similar things have been raised: http://indico.cern.ch/conferenceTimeTable.py?confId=12807 A.O.B. ====== none

11:00 → 11:05 Introduction

Speaker: Clemens Koerdt

11:05 → 11:25 Recent developments

- last COD meeting in Stockholm: http://indico.cern.ch/conferenceDisplay.py?confId=12759

Speakers: Christian Peter, Clemens Koerdt

11:25 → 11:45 Organizational matters

- organisation/availability of duty shifts - representation of COD Dech in Mondays Operations meetings - participation in COD working groups and related activities - participation in future COD meetings

Speaker: Clemens Koerdt, Victor Penso

11:45 → 11:50 A.O.B.