MinutesNov9

Introduction

Minutes of the Integration meeting, November 9, 2006

Attending

  • David Meyers, Rob Gardner, Jeff Porter, Shreyes Choila, Showen and GROW team, Steve Timm, Burt, Horst, Karthik, Doug, Illya Narky, Alain Roy, Gabriele, Xin Zhao, Alan Sill.

Agenda

  • Welcome Jeff Porter and Shreyas Cholia /LBNL
  • GIP input options subcommittee report - Steve Timm
  • ITB 0.5.1 SiteValidationTableITB051
  • AOB

GIP subcommittee report and discussion - Steve Timm

  • Steve reported from last week's phone meeting.
  • Anand advised that the configure GIP script is already doing the configure of SRM okay - if you know what to tell it. All that is needed is better docementation and perhaps better prompting - Steve will do this.
  • Discussion about publishing installed application release versions. There is a file in the $APP directory can be used by VO software installers to report which named releases they have installed. This file is read by a GIP script and fills an LDAP attribute. Question is whether this is dynamic or not, that is, does the site admin need to run something each time an application is updated or installed. Does it periodically update/check? (Steve checked it is in fact dymanic).
  • Gabriele: which glue schema attributes are used? Answer was gluehostapplicationsoftwareruntimeenvironment, or some such. Are we using it correctly? We did not resolve this here.
  • Showen volunteered to at how ATLAS releases are published in LCG, and make sure this is being done consistently in OSG (an WLCG requirement). Will send report to osg-int.

ITB 0.5.1

Review the status of ITB 0.5.1 validation, SiteValidationTableITB051 - David Meyer.
  • David notes that a CEMon status column has been created and is being filled by Gabriele. And that there has been much activity over the past week with GridCat, srmcp, VORS and LIGO validations.
  • Site reports:
    • BNL/Xin: A Condor configuration problem was fixed and the scheduler should now be okay. The table is not up to date for this site. Xin reports serious problems with the LDAP server, causing high load and instabilities. We discussed this. This is because the current installation uses an old LDAP version in VDT, not a newer version that comes with most new Linux installations. (Burt has previously reported on this in this meeting.) There is no plan in VDT to update this, since there is on-going work to use CEMon instead, with CEMon acting as the site level information service, with reports going to a BDII server at the GOC. Note that as the GOC has given ITB site names to EGEE, it was being queried by LCG resource brokers. RobQ? will remove ITB sites from that list.
    • FNAL ITB sites/Steve: no issues.
    • Iowa-GROW sites: beginning work on validation.
    • IUPUI:/Rob - no issues. Is testing new VOMS so the site is sometimes offline.
    • ITB-CIT-LIGO/David: focus is on config of BDII.
    • PSU/Murali: David is in contact with Murali - whose time is limited for ITB testing. Probably will remove this site.
    • OUHEP/Horst: No problems. cemon not publishing according to gabriele. they will resolve offline.
    • Purdue: remove, inactive.
    • TTU/Alan: this is coming up today on new hardware. Question about hooking to a backend production cluster, with a special ITB-queue. We don't see a problem with that.
  • Validation tasks:
    • Burt send mail that recipe for BDII:

From: Burt Holzman 
Date: November 9, 2006 3:28:20 PM CST
To: OSG-INT@OPENSCIENCEGRID.ORG
Subject: BDII Recipe

Rob,

This is Ransom's recipe for installing BDII.  Under this configuration it does not query the GRIS but runs the local information providers directly:

http://grow.its.uiowa.edu/infrastructure/cms/bdii.shtml	

I also have an inferior version which queries the GRIS that I can send out for interested parties.

- B
    • Burt will continue tests of srmclient - will send email with results. authz probs at some sites. some have not installed wn-client. expired crls too. osg grid variable not set right.
    • vors/robq - all okay as of last week. does site-verify tests. sometimes succeeds where gridcat fails. gridcat timeouts shorter.
    • gridcat - results since last week not updated in table.
    • Anand will do bdii/gip - Ransom reports.
    • Keith Chadwick's ITB validation process - failing at a number of sites. steve will update.
    • CEMon - Gabriele - are any additional sites planning to turn it on? yes: chicago will, ttu will. gabriele looks into rss collector. checks class ads. will send email. here it is:
From: Gabriele Garzoglio 
Date: November 9, 2006 3:42:05 PM CST
To: Robert Gardner 
Subject: CEMon command to check if a site reports to ReSS

The command is
condor_status -pool osg-ress-1 -constraint 'GlueCEInfoHostName == ""'

For example:

bash-2.05b$ condor_status -pool osg-ress-1 -constraint 'GlueCEInfoHostName == "cmssrv09.fnal.gov"'
Name          OpSys       Arch   State      Activity   LoadAv Mem   ActvtyTime
cmssrv09.fnal [?????????] [????] [????????] [???]  [??]   [Unknown]
cmssrv09.fnal [?????????] [????] [????????] [???]  [??]   [Unknown]
cmssrv09.fnal [?????????] [????] [????????] [???]  [??]   [Unknown]
cmssrv09.fnal [?????????] [????] [????????] [???]  [??]   [Unknown]
cmssrv09.fnal [?????????] [????] [????????] [???]  [??]   [Unknown]
                    Machines Owner Claimed Unclaimed Matched Preempting
              Total        0     0       0         0       0          0
                   (Omitted 31 malformed ads in computed attribute totals)
bash-2.05b$ condor_status -pool osg-ress-1 -constraint 'GlueCEInfoHostName == "fake.site.fqdn"'
bash-me2.05b$

Cheers
Gabriele
    • cdf - gabriele compastella(sp) will be validating itb sites. Steve will validate a config for squid. Not installed by default, though is part of the vdt. squid consumes a lot of memory, alain thought most site admins wouldn't want it on their gatekeepers, probably right. Alan mentions again whether this could be used for updating certificates for worker nodes. this has not been investigated.
    • atlas - xin reports torre will do a validation; new type of atlas pilot being developed.
  • Still need a new ValidationPage - Rob will update before next week.

AOB

  • None.

Major updates:
-- RobGardner - 08 Nov 2006

Topic revision: r5 - 16 Dec 2008 - 16:16:00 - KyleGross

Hello, TWikiGuest!
Register

 
TWIKI.NET

TWiki | Report Bugs | Privacy Policy

This site is powered by the TWiki collaboration platformCopyright by the contributing authors. All material on this collaboration platform is the property of the contributing authors..