MinutesVTBSep18

Introduction

• Attendees: Jeff, RobG, Steve Timm, Neha, RobQ, Arvind, Abhishek, Chris, Xin, Burt, Charles Huang, Anand, Alain
• Apologies: none
• Coordinates: Tuesday, 3:30pm Central; 510-665-5437, #1212
• Previous meetings, MeetingMinutes

ITB 0.7.0 background

• ReleaseDocumentation
• SiteValidationTable

VTB cycle for VDT 1.8.1 (Suchandra)

• Status of VTB cache - all ready . Basic install is working.
• Sites participating: UC, LBL, CIT-Charles
• Client install available too.

Shortcomings in service validation table

• There is not enough detail in ValidateYourSystemInVORS to be useful. Example: which tests are supposed to be check for success? Check VO support matrix. Check BDII information. All sites are passing. Not sure of status of PSU.
• There is not enough information in the ValidateBDII to be useful. Which tests succeed (provide example of successful ldap query output)? Are CE and SE critical attributes reported correctly? URLs are provided, but what is important to check there? To be done by small info service team
• There is not enough detail in ValidateRessAndCemon to be useful. Obviously a pre-requisite to run the valdation is an exisiting condor client (not mentioned). What should this data look like (provide an example)? The "web based representation" - what is being checked in detail? To be done by small info service team
• ValidateGratia - should be clarified. Is it enough to see that Gratia has contacted the site? Are instructions for Gratia collector relevant for all sites? Also providing a new reporting view for the ITB. Chris .
• ValidateGramWebServices - instructions are too thin for the Condor-G submission test. Can't we bring forward some of the WS-gram testing scripts? Have asked Charles Bacon
• ValSysLogNG - instructions are incomplete. Should not be logging to /tmp. Need to add php-syslog-ng based validation checks. Suchandra
• ValidateClients - this is weak and needs input. Jeff
• ValidateSquid - can we add more detail with explicit examples (for the first part about setting the http_proxy environment to the squid proxy host). Dave/Fnal working on this .
• ValidateGlexecAtSite - does not exist; there was an outline of instructions from Igor sent around. Steve working on this

VORS validation status

• Grid scan: http://scan.grid.iu.edu/cgi-bin/show_results?grid=2 (needs to be pruned to 13 ITB sites)
• VORS ITB view
• Why are "resource access points" available for some sites but not others?
• gk.phys.sinica.edu.tw:2119 should be removed from the table
• drop uct2dev from gridscan.

ReSS validation status

• The ReSS is supposed to capture the CE critical attributes, as specified https://twiki.grid.iu.edu/twiki/bin/view/Main/GIPAttributes.
• Are all sites reporting into https://osg-ress-4.fnal.gov:8443/cgi-bin/main.cgi correctly?

BDII validation status

• The BDII is supposed to capture both CE and storage critical attributes, as specified https://twiki.grid.iu.edu/twiki/bin/view/Main/GIPAttributes
• Most ITB sites are failing this, according to http://gip-validate.grid.iu.edu/itb/
• http://is-itb2.grid.iu.edu/cgi-bin/status.cgi - are critical attributes being reported correctly? Why are not all sites reporting?

RSV validation status

• Why have so few sites checked off in the validation table? Not sure why. OU and UCSD are two sites that are reporting okay. Will work with Murali.
• Sites: IUPUI-ITB - will not apply. LBNL: will work on 1.8.1 version. OU - Karthik will update, all is okay.
• http://peart.ucs.indiana.edu/docs/osg/vdt-package.html#test_output
• New fixes are in VDT 1.8.1, and an updated status page.
• Reporting to WLCG - work in progress.

LCG interoperability status

• No sites have checked off in the validation table - is this intentional? Burt and Tony will be checking this off.
• Have all sites correctly applied the patch distributed by Burt?
  • Tony will start doing some hello-world tests
• For SE's - no updates from LCG ("what is a validated SE"). May not make it in this ITB cycle. We will need some expert help at IU - Rob Q has been looking for errors. For example, OU applied the patch correctly but for some reason its showing up but the information is not visible from an ldap query. No info in the logfiles. Potentially serious issue.
• Are there any additional actions required by the ITB site admins? Hold off for right now.
• Have configure-gip patches made it into VDT 1.8.1 (Alain thinks Will has provided this.) Burt will check this tonight, from John W's auto-install.
• Revisit on Thursday.

glexec validation status

• What course to take given Steve's findings?
• Was broken but Igor has a work-around. Has a new glexec twiki page, and a validation page. Working on installing on his cluster.
• Not clear which of these fixes can make it into VDT 1.8.1 (not much time).
• Fix is to put a LD lib path into the glexec config file. There is a fair amount of post-installation configuration. How much can we simplify? Will come in an update.
• Also - glexec is available as rpms.
• Xin will also try this out at fermilab.
• Would be nice to have a full glide-in available.

Squid validation status

• We need one or two sites to step forward and test this.
• Dave Dykstra will be working on this.
• Will need to modify WN client to put fetch-crl in front of this - Steve.
• OU will volunteer as a site for squid.

Gratia validation status

• Where is ITB instance? ans: http://gratia-osg.fnal.gov:8881/gratia-reporting/
• Why is IPAS_OSG, Harvard, others listed?
• Are the 13 ITB sites correctly reporting?
• Can we easily get a snapshot of VO usage?
• Chris will provide a query which will give information about unknowns. For example, a osg-user-vo map problem.
• Condor will come with managed-fork, but its optional. But it must be turned on with vdt-control. Note one condor probe gets both.
• The RSV probes pull in a development condor, unrelated.

Storage validation status (dCache)

• Where are the detailed instructions for this? for site admins/for users - Alex S has indicated he will provide instructions
• Where are testing results being reported, eg. http://datagrid.lbl.gov/ has no information about OSG ITB 0.7 storage element test results

Storage validation status (Bestman)

• Where are the detailed instructions for this? for site admins/for users - Alex S has indicated he will provide instructions
• Where are testing results being reported, eg. http://datagrid.lbl.gov/ has no information about OSG ITB 0.7 storage element test results

Hello (fermi) grid results

• http://fermigrid.fnal.gov/monitor/fermigrid0-fermilab-vo-integration-monitor-detail.html
• Abhishek - notes that there have been days where all sites have been green - both Steve and Keith have been checking.

Grid exerciser

• The UC site was failing - space was an issue. Suchandra is working on resolving the issue - not sure what the current issues are.
• Charles will follow-up with site admins.

-- RobGardner - 17 Sep 2007