Summary

Description

This release was mainly focused on the introduction of SAM Operational Tools Monitoring. In addition we worked on bug fixing identified during the wide deployment of SAM Update 19.

Technical details:

• 70 tickets resolved
• Topology aggregation:
  • Added sanity check to compare differences between services declared in local ATP with central ATP
  • Enabled VOMS CSRF support
  • Fixed problem of invalid json output provided by ATP PI, which was affecting NCG
• Profile Management:
  • Unit test moved to Django 1.3
• MyWLCG changes:
  • MyWLCG Error Handling improved
  • Status view improved
  • Added monthly reports to the central MyEGI
  • Two new reports for T0/T1 sites
• Nagios configuration:
  • Enabled defining contacts for services
  • ncg.localdb partially migrated to ncg-metric-config (json format)
  • Optimized Java truststore generation
  • Added option to give permissions to run Nagios commands to anybody with a valid D
• Probes:
  • grid-monitoring-probes-ch.cern.sam:
    • MrsCheckDBInsertsDetailed probe improved
    • SamCheckUpdate probe improved
    • Added SAMCentralWebAPI probe
  • hr.srce.GridProxy-Get generates proxies with configurable lifetime
  • Probe libraries ported to SL6 (perl-TOM, python-GridMon and perl-GridMon)
  • Fixed mta-simple problem in grid-monitoring-probes-org.sam
• MRS metrics disabled on SAM/Nagios nodes
• SAM configuration changes (glite-yaim-nagios):
  • Consolidated/minimized number of httpd actions
  • Consolidation of YAIM variables (names uppercase)
  • OPS-MONITOR established as new SAM/Nagios configuration (for monitoring operational tools)
  • EGI report: switched Availability and Reliability labels
  • Decommission of MDDB
  • Improved MySQL database dump
  • Source code documentation
  • Removed dependencies on DAG repository

Package List

SAM-Nagios

 atp-1.26.4-1.el5
 glite-yaim-nagios-1.9.18-1.el5
 grid-monitoring-config-gen-0.91.3-1.el5
 grid-monitoring-probes-hr.srce-0.36.1-1.el5
 grid-monitoring-probes-org.sam-0.5.8-1.el5
 grid-monitoring-probes-ch.cern.sam-1.6.13-1.el5
 mrs-1.7.35-1.el5
 mywlcg-1.5.1-5.el5
 mywlcg-atp-api-1.26.1-1.el5
 mywlcg-atp-web-1.26.2-1.el5
 ncg-metric-config-1.2.3-1.el5
 perl-GridMon-1.0.62-1.el5
 python-GridMon-1.1.12-1.el5

SAM-Gridmon

ace-0.2.5-1.el5
 atp-1.26.4-1.el5
 mrs-1.7.35-1.el5
 mywlcg-1.5.1-5.el5
 mywlcg-atp-api-1.26.1-1.el5
 mywlcg-atp-web-1.26.2-1.el5
 glite-yaim-nagios-1.9.18-1.el5

Configuration Changes

Common

• New Yaim configuration variables:

  Component	Name	Description	Default	Mandatory	Example
  ATP	MSG_DEST_OSG_DOWNTIME	Messaging queue for OSG downtimes	Yes	Yes	/topic/grid.management.downtime.RSV

SAM-Gridmon

• New Yaim configuration variables

  Component	Name	Description	Default	Mandatory	Example
  MyWLCG	MYWLCG_TRENDS	Enable/disable Trends on MyWLCG	No	No	false

SAM-Nagios

• New Yaim configuration variables:

  Component	Name	Description	Default	Mandatory	Example
  nagios	NAGIOS_ENABLE_ANY_DN	Enable/disable of any DN to run Nagios commands on VO-nagioses	Yes	No	false
  NCG	CRO_BROKER_PASS	BROKER_PASSWORD value for msg.cro-ngi.hr broker	No	No	MyPass
  NCG	EGI1_BROKER_PASS	BROKER_PASSWORD value for egi-1.msg.cern.ch broker	No	No	MyPass
  NCG	EGI2_BROKER_PASS	BROKER_PASSWORD value for egi-2.msg.cern.ch broker	No	No	MyPass
  NCG	GR_BROKER_PASS	BROKER_PASSWORD value for broker.afroditi.hellasgrid.gr broker	No	No	MyPass
  NCG	NCG_SERVICE_NOTIFICATIONS_OPTIONS	Nagios notification options for services	No	No	u,c
  NCG	NCG_HOST_NOTIFICATIONS_OPTIONS	Nagios notification options for hosts	No	No	'cn'
  NCG	NCG_PROXY_LIFETIME	Defines the lifetime option to refresh_proxy probe in grid-monitoring-probes-hr.srce. If unset, probe will use default value 12. (Make sure that MyProxy server supports defined credential lifetime, otherwise the probe will fail)	No	No	12

Localdb changes

• Migration of localdb metric configuration to JSON config

  JSON config files stored in /etc/ncg-metric-config.d/ must use .conf suffix. Otherwise NCG will ignore them.

Modification and overriding of global values NCG localdb format:

# add/modify configuration parameter
  MODIFY_METRIC_CONFIG!metric!config!value
  # add/modify dependency
  MODIFY_METRIC_DEPENDENCY!metric!dep!value
  # add/modify attribute
  MODIFY_METRIC_ATTRIBUTE!metric!attr!value
  # add/modify parameter
  MODIFY_METRIC_PARAMETER!metric!param!value
  # add/modify flag
  MODIFY_METRIC_FLAG!metric!flag!value

should be done by using JSON configuration files in directory /etc/ncg-metric-config.d/:

# cat /etc/ncg-metric-config.d/modify.conf
{
   "metric" : {
      "attribute" : {
         "attr" : "value"
      },
      "parameter" : {
         "param" : "value"
      },
      "flags" : {
         "flag" : "value"
      },
      "config" : {
         "config" : "value"
      },
      "dependency" : {
         "dep" : 1
      },
   }
}

• Enable contacts for service flavour on a given host
  Use NCG localdb to add/enable contacts for all metrics associated to a given service flavour:

  ADD_SERVICEFLAVOURCONTACT!host!ServiceFlavour!email@email.com
  # enables contact even if NCG_ENABLE_NOTIFICATIONS is 0
  ENABLE_SERVICEFLAVOURCONTACT!host!ServiceFlavour!email@email.com

Known Issues

The packages sam-nagios-1.20.0-1.el5.noarch.rpm and sam-release-1.20.0-1.el5.noarch.rpm are not in the EGI repository (therefore setting a wrong version in /etc/sam-release for those using the EGI repo).

For machines running latest version of glite-UI (3.2.10-1 or higher): Please restart Nagios after yaim execution. Otherwise you may see problems similar to SAM-1693. service nagios restart

Upgrading a node with yum requires a package exclusion, e.g.: on sam-nagios yum update --exclude sam-gridmon on sam-gridmon yum update --exclude sam-nagios

Metrics ch.cern.sam.MrsCheckDBInserts and ch.cern.sam.MrsCheckDBInsertsDetailed have to be disabled manually. Please add following lines to file /etc/ncg/ncg.localdb REMOVE_METRIC!ch.cern.sam.MrsCheckDBInserts REMOVE_METRIC!ch.cern.sam.MrsCheckDBInsertsDetailed

Tickets List