Configuring RSV OSG Resource

1
Configuring RSV OSG Resource Service Validation

• Thomas Wang
• Grid Operations Center (OSG-GOC)
• Indiana University

2
RSV Overview

• RSV provides a flexible yet powerful
  infrastructure to validate resources and services
• Validation tests are performed via a series of
  simple probes
• Individual scripts in Perl, as well as worker
  scripts in other languages
• Perl Module with probe functions and basic tests
• These probes use Condor-cron as a scheduling
  infrastructure developed by the VDT
• Condor submission script for scheduling probes
• Condor consumer script will run a python script
  to upload RSV data to a central collector (Gratia
  Database)
• Data is collected with a Gratia based collector
  and stored in a MySQL database developed by the
  Gratia Group
• Provides central monitoring and long term storage
  of RSV Data
• The RSV Collector provides a central collection
  and storage point, while each monitoring host
  also provides a localized consumer page for the
  resources they monitor
• Probe metrics conform to WLCG specification and
  allow interoperability with SAM

3
Installing and Configuring RSV

• Getting the RSV Package
• The OSG package is available through VDT 1.8.1
  (OSG Prod CE 0.8.0 and ITB CE 0.7.1)
• Package includes Condor 6.9.4, a configuration
  and submission system for the probes, a mechanism
  to add new probes/alter configuration on existing
  probes, and a consumer system that displays probe
  output to a local HTML page and also uploads
  probe output data to the RSV collector
• Note to use an existing Condor 6.9.x install,
  before you begin installing your CE
• export VDTSETUP_CONDOR_DEVEL_LOCATION/my/condor-6
  .9.x/location/
• Note to install on a separate monitoring host
  you can get the RSV package with pacman
• pacman -get http//vdt.cs.wisc.edu/vdt_181_cacheO
  SG-RSV
• Configuring Metrics
• A sample_metrics.conf file is provided on the web
  (RSV Web pages and in your VDT Installation at
  VDT_LOCATION/osg-rsv/config/), with recommended
  time-intervals for various metrics.
• Copy VDT_LOCATION/osg-rsv/config/sample_metrics.c
  onf to VDT_LOCATION/osg-rsv/config/ltFQDN_of_CEgt_m
  etrics.conf, and edit the probe scheduling
  template
• If you have a metrics.conf or ltFQDN_of_CEgt_metrics
  .conf file from a previous RSV configuration,
  then you can just copy that file over to
  VDT_LOCATION/osgrsv/config/ltFQDN_of_monitored_hos
  tgt_metrics.conf
• It is still recommended to set the time-intervals
  for various metrics similar to the
  sample_metrics.conf template
• You can configure each host on a per-probe basis
  with the following format in the conf file
• Enable(on/off) Probe_at_Metric Cron Min,Cron Hr,Cron
  DoM,Cron Mon,Cron DoW

4
Configuration and Startup

• Configure OSG RSV
• Ensure that you have a Unix username that RSV can
  use to function, and an appropriate proxy owned
  by that user!
• VDT_LOCATION/vdt/setup/configure_osg_rsv
• --uri ltFQDN of host to probe1gt ltFQDN of host
  to probe2gt\
• --user ltRSV_USERgt --proxy /tmp/x509up_ultUSER_IDgt
  \
• --init --server y \
• --probes --gratia --verbose
  --print-local-time
• Enabling Gratia collection. Production sites use
  rsv.grid.iu.edu8880 ITB sites use
  rsv-itb.grid.iu.edu8880 as the value for the
  --report-to argument.
• VDT_LOCATION/vdt/setup/configure_gratia --probe
  metric \
• --report-to ltAppropriate_RSV_collector
  portgt
• If you would like RSV to enable VDT Apache to
  serve local site-level RSV status pages on the
  web
• VDT_LOCATION/vdt/setup/configure_osg_rsv
  --setup-for-apache
• Starting Condor-devel and RSV
• vdt-control --on condor-devel
• vdt-control --on osg-rsv
• If you configured RSV to use the VDT Apache to
  serve local status pages, then enable apache and
  turn it on as well
• VDT_LOCATION/vdt/setup/configure_apache --server
  y
• vdt-control --on apache

5
Notes and References

• Validation and Logs
• If enabled, the web status pages will be stored
  by default in VDT_LOCATION/osg-rsv/output/html/st
  atus.html
• These pages can be viewed locally or via the web
  at http//ltHost's FQDNgt8443/rsv/
• Note User must have their OSG Approved Cert
  loaded into their web browser
• RSV Logs are kept in VDT_LOCATION/osg-rsv/logs/
• Check the probes folder for probe error logging
• Can check consumer folder for logging of Gratia
  uploading and the HTML status page
• Changes to RSV
• Any changes to the .conf file or the .spec files
  will require that your stop RSV, run its
  configure script again, and then start it
• vdt-control --off osg-rsv
• configure_osg_rsv for the probes, as shown
  previously
• vdt-control --on osg-rsv
• Official Documentation Found at
  http//rsv.grid.iu.edu/documentation
• Comprehensive Install and Configuration Guide
• Quick Install and Configuration Guide
• Readme and Probes Tarball (probes only)
• RSV Sample Status Page
• Web Interface for Grtaia Collector (coming soon)

6
Discussion

• Discussion Items
• Bug Status
• Print Local Time (fixed in VDT 1.8.1d)
• Classad bug with Condor (should be fixed in next
  VDT release)
• HTML Consumer Bug (should be fixed in next VDT
  release)
• Proxies in RSV (Alternatives and Security)
• Current proxy setup
• MyProxy, Robot Certs, other suggestions
• Discuss with OSG Security Group
• Future of Condor-cron in RSV
• Lessons learned from ReSS Classad Probe (conflict
  with existing condor instances)
• RSV Version 2
• Project Plan is being developed
• Suggestions for Version 2 Probe set
• Suggestions and Volunteers for writing new probes
  
• Any other thoughts, comments, suggestions?
• Contact the RSV Development Team rsv-dev at
  opensciencegrid.org
• Special Thanks to Arvind for RSV reference slides
  and Documentation from the last Site Admin Meeting