MinutesRSVNagiosDec21

Introduction

• Background
  • https://twiki.cern.ch/twiki/bin/view/LCG/GridServiceMonitoringInfo
  • Recent service reliability workshop at CERN, http://indico.cern.ch/conferenceDisplay.py?confId=20080
• Attending: US ATLAS: Rob, Tomasz, John; OSG: Rob Quick, Tom Wang, Arvind Gopu
• Phone: (605) 475-6000, Access code: 735188; Dial 6 to mute/un-mute.

RSV to US ATLAS Nagios

• Arvind thinks best way is to push data from RSV to BNL.
• There is a GMS publish script provided by James Casey. Is used by the WLCG SAM instance.
• Getting data to BNL menas there are firewall issues; proposal was to put a middle=man Mysql server to publish RSV data to.
• Tomasz would prefer that OSG publish data in a web-visible area, that can be parsed remotely.
• RSV team would like to limit the number of publishing methods.
• John: ATLAS would like to pull rather than push. For pushing, more infrastructure is required on the receiving end.
• RSV team is concerned about about the policy issues.
• Arvind: what is needed is a common format specification and common authentication sytem.
• Regarding the format - we might follow XML exchange format standard from WLCG.
• Regarding policy, Rob Q will write a policy document for RSV probes and will request comment.
• Can we make progress with non-sensitive probes? (All the ATLAS monitoring information at the moment is non-sensitive, publically visible)
• RSV team will discuss and will try to provide a pull-method interim solution w/ non-senstive data.

Site level RSV to Nagios

• The RSV team has already been working w/ WLCG on this approach. There's a development version running by James Casey. There's also discussion about how to set this up at the site level, but the configuration is difficult. Arvind will forward a url.
• Tomasz has done something similar. Use passive monitoring feature of Nagios. The RSV probe has a wrapper that has an http client which allows it to upload the RSV probe result to Nagios. Tomasz will re-circulate his mail w/ code, etc.
• The RSV team will be evaluating both approaches as part of their RSV 2.0 work plan.
• Rob Q notes there will be Nagios development to replace current gridscan/VORS functionality at the GOC.

RSV GOC to WLCG SAM

• RSV team has been working on the transfer technology to publish RSV data to SAM. They currently are publishing ITB data.
• Arvind: we'll start publishing OSG 0.8 production data next week, but the SAM developers will have to modify the portal when they return from Christmas break.
• Notes that resources must be in BDII, but hostnames are needed as well as site names (may need to use OSG site registration database). We'll probably have to work this through by hand before OIM is developed/deployed since official reporting begins in January of 2008.
• James, Piotr, David - are points of contact on WLCG SAM side. Will likely need to iterate to get sitenames, hostnames attributed correctly. Also to exclude opportunistic OSG resources which are associated w/ US Tier2 centers but not subject to the MOU.
• Links
  • OSG: https://lcg-sam.cern.ch:8443/sam-uncert/sam.py?sensors=OSGCE&regions=OpenScienceGrid&vo=ops&order=SiteName&funct=ShowSensorTests
  • ITB: https://lcg-sam.cern.ch:8443/sam-pps/sam.py?sensors=OSGCE&regions=OpenScienceGrid&vo=ops&order=SiteName&funct=ShowSensorTests

-- RobertGardner - 21 Dec 2007