Site Validation Session Report

1
Site Validation Session Report

• Co-Chairs
• Piotr Nyczyk, CERN IT/GD
• Leigh Grundhoefer, IU / OSG
• Notes from Judy Novak
• WLCG-OSG-EGEE Workshop
• CERN, June 19-20th 2006

2
Service Availability Monitoring (SAM) -
extension of SFT

• generalized framework to monitor all LCG/EGEE
  services and not only CE BDII, RB, LFC, FTS,
  etc.
• most of the sensors run remotely (from central
  machine)
• no installation needed on service machines
• moved from MySQL to Oracle, optimized data schema

Available at https//lcg-sam.cern.ch8443/sam/sa
m.cgi
3

• SAM sensors
• currently BDII (Taiwan), RB (RAL), CE, SRM, LFC,
  FTS, SE (CERN)
• release updates SAM (SFT)
• certifying current tests with each new release
• Create update tests as necessary
• CA cert. releases are special
• Availability views
• current, daily, weekly, monthly
• For CE, SE, SRM, siteBDII
• displayed with GridView

http//glite.cvs.cern.ch/cgi-bin/glite.cgi/sft2/te
sts/
4
OSG Validation services

• CE/SE Validation aggregation VORS - site
  scanner, BDII info
• http//vors.grid.iu.edu/
• OSG VOs VOMS validation
• http//voms-monitor.grid.iu.edu/
• GridEX - application validation ( pilot job
  submissions )
• http//www.cs.wisc.edu/condor/tools/exerciser/
• Site Policy template and publication
• http//vors.grid.iu.edu/site_policies.html
• GIP Validation
• http//grow.its.uiowa.edu/osg-gip/Production.shtml
  
• Monitoring validation MonALisa Client status
  (VO Jobs I/O)
• http//grid02.uits.indiana.edu8080/stats?pagesum
  mary
• GridCat and the MIS-CI client
• http//osg-cat.grid.iu.edu/ - Production
  instance
• Client software http//software.grid.iu.edu/pacma
  n/tarballs/misci-0.4.1.tar.gz

5
Summary

• It seems to be impossible to avoid
  cross-monitoring (OSG monitoring doesn't include
  LCG-specific services, and the other way around)
• We should synchronize on VO level, but LCG/EGEE
  is also using regional structuring

6
OSG and EGEE Validation Interoperability

• Site discovery - using discovered sites using
  BDII
• Ops VO - supported only on OSG sites which are
  interoperable. (fully deployed in July)
• How can we determine if EGEE site is
  interoperable? Review certain BDII informations
• Cross installation of necessary tools and
  libraries for site validation
• LCG tools - added as optionally installed
  package for OSG sites
• OSG environment variables - ? (GIP)

7
OSG and EGEE Validation Interoperability (cont)

• Use of existing GGUS- OSG GOC ticket exchange for
  error reporting
• SAM database to use contact information for OSG
  GOC
• Issue of coordinating scheduled downtime
• OSG GOC will maintain a web page with downtimes
• Propose review of effort to add OSG specific
  validations to SAM framework.
• Testing and iterative development will be
  accomplished using Pre-Production sites and OSG
  ITB

8
DB monitoring in SAM for Tier 1s (Dirk
Duellmann)

• Jobs are connecting to the DB with either http
  (VO lib) or direct Oracle (instant client)
  
• Should be completed by October when experiments
  will start using DBs
• CMS Alice don't need them, but only 'squid
• existing DB monitoring is too detailed for
  SAM/SFT, but SAM could provide highlevel
  monitoring of DB service
• some DB services (like LFC) are already tested by
  SAM, BUT only the functionality is tested, not
  the DB! The test could be
• threshold for connection between T0 -gt T1
• user access (squid)
• client latency (?)
• Oracle client will be installed on the Worker
  Nodes

9
Comments/Discussion