LCG - BDII - Generic Configuration Reference

Document identifier:	LCG-GIS-CR-BDII
Date:	16 January 2006
Author:	LCG Deployment - GIS team;Retico,Antonio;Vidic,Valentin;
Version:

Abstract: Configuration steps done by the YAIM script 'configure_BDII'

Contents

• Introduction
• Variables
• Create EDG users
  • Specification of function: config_edgusers
  
  
• Set-up BDII
  • Specification of function: config_bdii
  
  
• Set-up GridICE agent
  • Specification of function: config_fmon_client
  
  
• Source code
  • config_edgusers
  • config_bdii
  • config_fmon_client
  
  
• About this document ...

Introduction

This document lists the manual steps for the installation and configuration of a LCG BDII Node.
Furthermore it provides a specification of the YAIM functions used to configure the node with the script-based configuration.

The configuration has been tested on a standard Scientific Linux 3.0 Installation.

Link to this document:
This document is available on the Grid Deployment web site

http://www.cern.ch/grid-deployment/gis/lcg-GCR/index.html

Variables

In order to set-up a BDII node, you need at least the following variables to be correctly configured in the site configuration file (site-info.def):

BATCH_LOG_DIR :

Your batch system log directory.

BDII_FCR :

Set the URL of the Freedom of Choice for Rescources URL.

BDII_HOST :

BDII Hostname.

BDII_HTTP_URL :

URL pointing to the BDII configuration file.

BDII_REGIONS :

List of node types publishing information on the bdii. For each item listed in the BDII_REGIONS variable you need to create a set of new variables as follows:

BDII_$<$REGION$>$_URL :

URL of the information producer (e.g.: BDII_CE_URL="URL of the CE information producer", BDII_SE_URL="URL of the SE information producer".

CE_BATCH_SYS :

Implementation of site batch system. Available values are ``torque'', ``lsf'', ``pbs'', ``condor'' etc.

CE_HOST :

Computing Element Hostname.

CRON_DIR :

Yaim writes all cron jobs to this directory. Change it if you want to turn off Yaim's management of cron.

GRIDICE_SERVER_HOST :

GridIce server host name (usually run on the MON node).

INSTALL_ROOT :

Installation root - change if using the re-locatable distribution.

MON_HOST :

MON Box Hostname.

MY_DOMAIN :

site's domain name.

SITE_NAME :

Your GIIS.

USERS_CONF :

Path to the file containing a list of Linux users (pool accounts) to be created. This file should be created by the Site Administrator, which contains a plain list of the users and IDs. An example of this configuration file is given in /opt/lcg/yaim/examples/users.conf.

VOS :

List of supported VOs.

Create EDG users

Author(s): Retico,Antonio
Email : support-lcg-manual-install@cern.ch

This chapter describes the configuration steps done by the yaim function 'config_edgusers'.

Many of the services running on LCG service nodes are owned by the user edguser. The user edguser belongs to the group edguser and it has got a home directory in /home.

The user edginfo is required on all the nodes publishing information on the Information System. The user belongs to the group edginfo and it has got a home directory in /home.

No special requirements exists for the ID of the above mentioned users and groups.

The function creates bothedguser and edginfo groups and users.

• group edguser: the group is created with group ID 995.
• user edguser: the user is created with group ID 995 and its home is /home/edguser.
• group edginfo: the group is created with group ID 999.
• user edginfo: the user is created with group ID 999 and its home is /home/edguser.

Specification of function: config_edgusers

The function 'config_edgusers' needs the following variables to be set in the configuration file:

INSTALL_ROOT :

Installation root - change if using the re-locatable distribution.

USERS_CONF :

Path to the file containing a list of Linux users (pool accounts) to be created. This file should be created by the Site Administrator, which contains a plain list of the users and IDs. An example of this configuration file is given in /opt/lcg/yaim/examples/users.conf.

VOS :

List of supported VOs.

The original code of the function can be found in:

/opt/lcg/yaim/functions/config_edgusers

The code is reproduced also in 6.1..

Set-up BDII

Author(s): Vidic,Valentin
Email : support-lcg-manual-install@cern.ch

This chapter describes the configuration steps done by the yaim function 'config_bdii'.

This functions is used to configure two types of BDIIs: site BDII and top BDDI.

At each site, a site BDII collects information about all resources present at a site (i.e. data from all GRISes of the site).

A top BDII collects all information coming from site BDIIs and stores them in a permanent database.

BDII is configured in <INSTALL_ROOT>/bdii/etc/bdii.conf. For top BDII the configuration is:

BDII_PORT_READ=2170
BDII_PORTS_WRITE="2171 2172 2173"
BDII_USER=edguser
BDII_BIND=mds-vo-name=local,o=grid
BDII_PASSWD=<random_password>
BDII_SEARCH_FILTER='*'
BDII_SEARCH_TIMEOUT=30
BDII_BREATHE_TIME=60
BDII_AUTO_UPDATE=yes
BDII_AUTO_MODIFY=no
BDII_DIR=<INSTALL_ROOT>/bdii/
BDII_UPDATE_URL=<BDII_HTTP_URL>
BDII_UPDATE_LDIF=http://
SLAPD=/usr/sbin/slapd
SLAPADD=/usr/sbin/slapadd

BDII is setup with the base DN of mds-vo-name=local,o=grid and automatic generation of <INSTALL_ROOT>/bdii/etc/bdii-update.conf by downloading from <BDII_HTTP_URL>. BDII password is generated using mkpasswd or from RANDOM shell variable.

For local BDII the configuration is:

BDII_PORT_READ=2170
BDII_PORTS_WRITE="2171 2172 2173"
BDII_USER=edguser
BDII_BIND=mds-vo-name=<SITE_NAME>,o=grid
BDII_PASSWD=<random_password>
BDII_SEARCH_FILTER='*'
BDII_SEARCH_TIMEOUT=30
BDII_BREATHE_TIME=60
BDII_AUTO_UPDATE=no
BDII_AUTO_MODIFY=no
BDII_DIR=<INSTALL_ROOT>/bdii/
BDII_UPDATE_URL=<BDII_HTTP_URL>
BDII_UPDATE_LDIF=http://
SLAPD=/usr/sbin/slapd
SLAPADD=/usr/sbin/slapadd

Base DN is set to mds-vo-name=<SITE_NAME>,o=grid. <INSTALL_ROOT>/bdii/etc/bdii-update.conf is created from the values of variables <BDII_REGIONS> and <BDII_<region>_URL>.

LDAP schemas are configured in <INSTALL_ROOT>/bdii/etc/schemas:

/etc/openldap/schema/core.schema
/opt/glue/schema/ldap/Glue-CORE.schema
/opt/glue/schema/ldap/Glue-CE.schema
/opt/glue/schema/ldap/Glue-CESEBind.schema
/opt/glue/schema/ldap/Glue-SE.schema
/opt/lcg/schema/ldap/SiteInfo.schema

BDII service is restarted and configured to start on boot.

Specification of function: config_bdii

The function 'config_bdii' needs the following variables to be set in the configuration file:

BDII_FCR :

Set the URL of the Freedom of Choice for Rescources URL.

BDII_HOST :

BDII Hostname.

BDII_HTTP_URL :

URL pointing to the BDII configuration file.

BDII_REGIONS :

List of node types publishing information on the bdii. For each item listed in the BDII_REGIONS variable you need to create a set of new variables as follows:

BDII_$<$REGION$>$_URL :

URL of the information producer (e.g.: BDII_CE_URL="URL of the CE information producer", BDII_SE_URL="URL of the SE information producer".

INSTALL_ROOT :

Installation root - change if using the re-locatable distribution.

SITE_NAME :

Your GIIS.

The original code of the function can be found in:

/opt/lcg/yaim/functions/config_bdii

The code is also reproduced in 6.2..

Set-up GridICE agent

Author(s): Retico,Antonio
Email : support-lcg-manual-install@cern.ch

This chapter describes the configuration steps done by the yaim function 'config_fmon_client'.

The LCG nodes can produce data for the GridICE monitoring system.
The data are then sent to a collector server node which will then be queried by the LCG central GridICE monitoring service.

If you are running agents on the nodes (data producers), you should also run a GridICE collector server to collect information from your agents.

In the default LCG-2 configuration the MON node runs the GridICE collector node.

Before going forward with configuration, please assure the following RPMs to be installed (they should have been distributed with the node RPMs).

edg-fabricMonitoring
edt_sensor

In order to enable GridICE agent on a LCG node:

• Create and configure the file /opt/edg/var/etc/edg-fmon-agent.conf as follows:

  # Sensor file for edg-fmonagent
  MSA
  
  
  
  
  	Transport
  
  
  
  		UDP
  			Server  <GRIDICE_SERVER_HOST>
  			Port 12409
  			FilterMetrics KeepOnly
  				11001
  				11011
  				11021
  				11101
  				11202
  				11013
  				11022
  				11031
  				11201
  				10100
  				10101
  				10102
  				10103
  				10104
  				10105
  
  
  
  
  	Sensors
  
  		edtproc 
  			CommandLine /opt/edt/monitoring/bin/GLUEsensorLinuxProc
  			MetricClasses
  				edt.uptime  
  				edt.cpu  
  				edt.memory  
  				edt.disk  
  				edt.network  
  				edt.ctxint  
  				edt.swap  
  				edt.processes  
  				edt.sockets  
  				edt.cpuinfo  
  				edt.os  
  				edt.alive  
  				edt.regfiles  
  
  		sensor1 
  			CommandLine $(EDG_LOCATION)/libexec/edg-fmon-sensor-systemCheck
  			MetricClasses
  				executeScript  
  
  
  	Metrics
  		11001
  			MetricClass edt.uptime 
  		11011
  			MetricClass edt.cpu 
  		11021
  			MetricClass edt.memory 
  		11101
  			MetricClass edt.disk 
  		11202
  			MetricClass edt.network 
  			Parameters
  				interface	eth0
  		11013
  			MetricClass edt.ctxint 
  		11022
  			MetricClass edt.swap 
  		11031
  			MetricClass edt.processes 
  		11201
  			MetricClass edt.sockets 
  		10100
  			MetricClass edt.cpuinfo 
  		10101
  			MetricClass edt.os 
  		10102
  			MetricClass edt.alive 
  		10103
  			MetricClass edt.regfiles 
  		10104
  			MetricClass executeScript 
  			Parameters
  				command	/opt/edt/monitoring/bin/CheckDaemon.pl --cfg /opt/edt/monitoring/etc/gridice-role.cfg
  		10105
  			MetricClass executeScript 
  			Parameters
  				command	/opt/edt/monitoring/bin/PoolDir.pl
  			
  
  
  	Samples
  		verylowfreq
  			Timing 3600 0
  			Metrics 
  				10100
  				10101
  		lowfreq
  			Timing 1800 0
  			Metrics 
  				11001
  		proc0
  			Timing 30 0
  			Metrics 
  				10102
  		proc1
  			Timing 60 0
  			Metrics 
  				11011
  				11021
  				11101
  				11202
  				11013
  				11022
  				11031
  				11201
  		proc2
  			Timing 300 0
  			Metrics 
  				10103
  				10105
  		proc3
  			Timing 120 0
  			Metrics 
  				10104
  

  WARNING: be very careful not to use <SPACE> characters to indent lines in this configuration file. Use <TAB> (or nothing) instead. The edg-fmon-agent does not allow spaces at the beginning of a row in the configuration file.
  

  The parameter <GRIDICE_SERVER_HOST> is the complete hostname of the node that runs the GridICE collector server and publishes the data on the information system. The collector node will have to run a plain GRIS for this.
  

  The information is sent to the collector node via UDP (port 12409).

• start the GridICE agent

  > chkconfig edg-fmon-agent on
  > service edg-fmon-agent stop
  > service edg-fmon-agent start
  

Specification of function: config_fmon_client

The function 'config_fmon_client' needs the following variables to be set in the configuration file:

BATCH_LOG_DIR :

Your batch system log directory.

CE_BATCH_SYS :

Implementation of site batch system. Available values are ``torque'', ``lsf'', ``pbs'', ``condor'' etc.

CE_HOST :

Computing Element Hostname.

CRON_DIR :

Yaim writes all cron jobs to this directory. Change it if you want to turn off Yaim's management of cron.

GRIDICE_SERVER_HOST :

GridIce server host name (usually run on the MON node).

INSTALL_ROOT :

Installation root - change if using the re-locatable distribution.

MON_HOST :

MON Box Hostname.

MY_DOMAIN :

site's domain name.

The original code of the function can be found in:

/opt/lcg/yaim/functions/config_fmon_client

The code is also reproduced in 6.3..

Source code

config_edgusers

config_edgusers(){

INSTALL_ROOT=${INSTALL_ROOT:-/opt}

check_users_conf_format

if ( ! id edguser > /dev/null 2>&1 ); then
    useradd -r -c "EDG User" edguser
    mkdir -p /home/edguser
    chown edguser:edguser /home/edguser
fi

if ( ! id edginfo > /dev/null 2>&1 ); then
    useradd -r -c "EDG Info user" edginfo
    mkdir -p /home/edginfo
    chown edginfo:edginfo /home/edginfo
fi

if ( ! id rgma > /dev/null 2>&1 ); then
    useradd -r -c "RGMA user" -m -d ${INSTALL_ROOT}/glite/etc/rgma rgma
fi

# Make sure edguser is a member of each group

awk -F: '{print $3, $4, $5}' ${USERS_CONF} | sort -u | while read gid groupname virtorg; do
    if ( [ "$virtorg" ] && echo $VOS | grep -w "$virtorg" > /dev/null ); then
	# On some nodes the users are not created, so the group will not exist
	# Isn't there a better way to check for group existance??
	if ( grep "^${groupname}:" /etc/group > /dev/null ); then
	    gpasswd -a edguser $groupname  > /dev/null
	fi
    fi
done

return 0
}

config_bdii

config_bdii(){
#
# Configures the BDII.
#
# If SITE_BDII=yes configures as a site BDII otherwise top level
#
# Uses CE_HOST SE HOST RB_HOST and PX_HOST.
#
# These values should be changed the common ones.
#

requires BDII_HOST

INSTALL_ROOT=${INSTALL_ROOT:-/opt}

mkdir -p $INSTALL_ROOT/bdii/etc
chown edguser $INSTALL_ROOT/bdii/etc
mkdir -p $INSTALL_ROOT/bdii/var
chown edguser $INSTALL_ROOT/bdii/var

BDII_AUTO_MODIFY=no

if ( ! echo "${NODE_TYPE_LIST}" | grep BDII > /dev/null ); then

	# We're a site BDII

	requires BDII_REGIONS BDII__URL

	BDII_BIND=mds-vo-name=$SITE_NAME,o=grid
	BDII_AUTO_UPDATE=no

	rm -f $INSTALL_ROOT/bdii/etc/bdii-update.conf
	for REGION in $BDII_REGIONS; do
            echo "$REGION  `eval echo '$'BDII_${REGION}_URL`" >> $INSTALL_ROOT/bdii/etc/bdii-update.conf
	done

else

	# We're a top level BDII

	requires BDII_HTTP_URL

	if [ "$BDII_FCR" ]; then
	    BDII_AUTO_MODIFY=yes
	fi

	BDII_BIND=mds-vo-name=local,o=grid
	BDII_AUTO_UPDATE=yes

fi

pass=`mkpasswd -s 0 2> /dev/null` || pass=$RANDOM

cat << EOF > $INSTALL_ROOT/bdii/etc/bdii.conf
BDII_PORT_READ=2170
BDII_PORTS_WRITE="2171 2172 2173"
BDII_USER=edguser
BDII_BIND=$BDII_BIND
BDII_PASSWD=$pass
BDII_SEARCH_FILTER='*'
BDII_SEARCH_TIMEOUT=30
BDII_BREATHE_TIME=60
BDII_AUTO_UPDATE=$BDII_AUTO_UPDATE
BDII_AUTO_MODIFY=$BDII_AUTO_MODIFY
BDII_DIR=$INSTALL_ROOT/bdii/
BDII_UPDATE_URL=$BDII_HTTP_URL
BDII_UPDATE_LDIF=${BDII_FCR:-http://}
SLAPD=/usr/sbin/slapd
SLAPADD=/usr/sbin/slapadd

EOF

cat << EOF > $INSTALL_ROOT/bdii/etc/schemas
/etc/openldap/schema/core.schema
/opt/glue/schema/ldap/Glue-CORE.schema
/opt/glue/schema/ldap/Glue-CE.schema
/opt/glue/schema/ldap/Glue-CESEBind.schema
/opt/glue/schema/ldap/Glue-SE.schema
/opt/lcg/schema/ldap/SiteInfo.schema

EOF

/sbin/chkconfig --add bdii
/sbin/service bdii stop
/sbin/service bdii start

return 0
}

config_fmon_client

config_fmon_client(){

# Modified by Cristina Aiftimiei (aiftim <at> pd.infn.it): 
# Modified by Enrico Ferro (enrico.ferro <at> pd.infn.it)
# 	host kernel version no more published
# 	user DN hidden by default
# 	job monitoring resource refresh for jobs in on Q/R status disabled by default
# 	support new job monitoring probe
# 	support new LRMSInfo probe

INSTALL_ROOT=${INSTALL_ROOT:-/opt}

requires GRIDICE_SERVER_HOST

mkdir -p ${INSTALL_ROOT}/edg/var/etc
> ${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg

# Job-Monitoring parameters
JM_TMP_DIR=/var/spool/gridice/JM
LAST_HOURS_EXEC_JOBS=2
mkdir -p ${JM_TMP_DIR}/new
mkdir -p ${JM_TMP_DIR}/ended
mkdir -p ${JM_TMP_DIR}/subject
mkdir -p ${JM_TMP_DIR}/processed

#  Monitoring of processes/daemon with gridice
if ( echo "${NODE_TYPE_LIST}" | grep CE > /dev/null ); then
cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
[ce-access-node]
gsiftp ^[\s\w\/\.-]*ftpd
edg-gatekeeper ^[\s\w\/\.-]*edg-gatekeeper
globus-mds ^[\s\w\/\.-]*${INSTALL_ROOT}/globus/libexec/slapd
fmon-agent ^[\s\w\/\.-]*fmon-agent
lcg-bdii-fwd    ^[\s\w\/\.-]*bdii-fwd
lcg-bdii-update ^[\w\/\.-]*perl\s[\s\w\/\.-]*bdii-update
lcg-bdii-slapd  ^[\w\/\.-]*slapd\s[\s\w\/\.\-]*bdii
EOF

if [ "$CE_BATCH_SYS" = "torque" ] || [ "$CE_BATCH_SYS" = "pbs" ] || [ "$CE_BATCH_SYS" = "lcgpbs" ]; then
   cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
pbs-server ^[\s\w\/\.-]*pbs_server
maui ^[\s\w\/\.-]*maui
EOF
fi
if [ "$CE_BATCH_SYS" = "lsf" ]; then
   cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
lsf-lim ^[\s\w\/\.-]*lim
lsf-pim ^[\s\w\/\.-]*pim
lsf-res ^[\s\w\/\.-]*res
lsf-sbatchd ^[\s\w\/\.-]*sbatchd
EOF
MASTER="`lsclusters |grep -v MASTER |awk '{print \$3}'`"
if [ "$CE_HOST" = "$MASTER" -o "$CE_HOST" = "$MASTER.$MY_DOMAIN" ]; then
       cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
lsf-mbatchd ^[\s\w\/\.-]*mbatchd
EOF
fi

fi
cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
[ce-access-node end]
EOF
fi

if ( echo "${NODE_TYPE_LIST}" | grep SE > /dev/null ); then
cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
[se-access-node]
gsiftp ^[\s\w\/\.-]*ftpd
globus-mds ^[\s\w\/\.-]*${INSTALL_ROOT}/globus/libexec/slapd.*2135.*
fmon-agent ^[\s\w\/\.-]*fmon-agent
[se-access-node end]
EOF
fi

if ( echo "${NODE_TYPE_LIST}" | grep SE_dpm_mysql > /dev/null ); then
cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
[dpm-master-node]
globus-mds ^[\s\w\/\.-]*/opt/globus/libexec/slapd.*2135.*
fmon-agent ^[\s\w\/\.-]*fmon-agent
dpm-master ^[\s\w\/\.-]*dpm
dpm-names ^[\s\w\/\.-]*dpnsdaemon
MySQL ^[\s\w\/\.-]*mysqld
srm-v1-interface ^[\s\w\/\.-]*srmv1
srm-v2-interface ^[\s\w\/\.-]*srmv2
gsiftp ^[\w,\/,-]*ftpd
rfio ^[\w,\/,-]*rfiod
[dpm-master-node end]
EOF
fi

if ( echo "${NODE_TYPE_LIST}" | grep SE_dpm_disk > /dev/null ); then
cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
[dpm-pool-node]
gsiftp ^[\w,\/,-]*ftpd
rfio ^[\w,\/,-]*rfiod
[dpm-pool-node end]
EOF
fi

if [ "X$GRIDICE_SERVER_HOST" = "X`hostname -f`" ]; then
cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
[gridice-collector]
gridice-mds ^[\s\w\/\.-]*${INSTALL_ROOT}/globus/libexec/slapd.*2136.*
fmon-server ^[\s\w\/\.-]*fmon-server
[gridice-collector end]
EOF
fi

if [ "X$MON_HOST" = "X`hostname -f`" ]; then
cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
[rgma-monbox]
ntpd ^[\s\w\/\.-]*ntpd
tomcat [\s\w\/\.-]tomcat
fmon-agent ^[\s\w\/\.-]*fmon-agent
[rgma-monbox end]
EOF
fi

if ( echo "${NODE_TYPE_LIST}" | grep RB > /dev/null ); then
cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
[broker]
ftp-server              ^[\s\w\/\.-]*ftpd
job-controller          ^[\s\w\/\.-]*edg-wl-job_controller
condor-master           ^[\s\w\/\.-]*condor_master
logging-and-bookkeeping ^[\s\w\/\.-]*edg-wl-bkserverd
condorg-scheduler       ^[\s\w\/\.-]*condor_schedd
log-monitor             ^[\s\w\/\.-]*edg-wl-log_monitor
local-logger            ^[\s\w\/\.-]*edg-wl-logd
local-logger-interlog   ^[\s\w\/\.-]*edg-wl-interlogd
network-server          ^[\s\w\/\.-]*edg-wl-ns_daemon
proxy-renewal           ^[\s\w\/\.-]*edg-wl-renewd
workload-manager        ^[\s\w\/\.-]*edg-wl-workload_manager
fmon-agent              ^[\s\w\/\.-]*fmon-agent
[broker end]
EOF
fi

if ( echo "${NODE_TYPE_LIST}" | grep BDII > /dev/null ); then
cat <<EOF >>${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
[bdii]
lcg-bdii-fwd    ^[\s\w\/\.-]*bdii-fwd
lcg-bdii-update ^[\w\/\.-]*perl\s[\s\w\/\.-]*bdii-update
lcg-bdii-slapd  ^[\w\/\.-]*slapd\s[\s\w\/\.\-]*bdii
fmon-agent      ^[\s\w\/\.-]*fmon-agent
[bdii end]
EOF
fi


# Configuration File for JobMonitoring
# If not definied before, use these defaults
GRIDICE_HIDE_USER_DN=${GRIDICE_HIDE_USER_DN:-yes}
GRIDICE_REFRESH_INFO_JOBS=${GRIDICE_REFRESH_INFO_JOBS:-no}

cat <<EOF >${INSTALL_ROOT}/gridice/monitoring/etc/JM.conf
##
## /opt/gridice/monitoring/etc/JM.conf
##

LRMS_TYPE=${CE_BATCH_SYS}

# --jm-dir=<$JM_TMP_PATH> (default /var/spool/gridice/JM) -- inside this directory
#                 will be created "new/" "ended/" "subject/" "processed/";
#                 when messlog_mon.pl is restarted it has to delete all
#                 "processed/.jmgridice*" files
JM_TMP_DIR=${JM_TMP_DIR}

# "--lrms-path=<LRMS_SPOOL_DIR>" (path for logs of batch-system)
LRMS_SPOOL_DIR=${BATCH_LOG_DIR}

# "--hide-subject=<yes|no>" (default: yes)
HIDE_USER_DN=${GRIDICE_HIDE_USER_DN}

# "--interval=<interval for ended jobs>", in hours (default: 2)
LAST_HOURS_EXEC_JOBS=${LAST_HOURS_EXEC_JOBS}

# <yes|no> (set the parameter "--no-update" if "no", otherwise no parameter is passed)
REFRESH_INFO_FOR_RUNNING_JOBS=${GRIDICE_REFRESH_INFO_JOBS}
EOF

# End configuration File for JobMonitoring

cat <<EOF >${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf
# template Sensor file for edg-fmonagent
# ** DO NOT EDIT **
# Generated from template: /usr/lib/lcfg/conf/fmonagent/sensors.cfg

MSA




	Transport



		UDP
			Server ${GRIDICE_SERVER_HOST}
			Port 12409
			FilterMetrics KeepOnly
				11001
				11011
				11021
				11101
				11202
				11022
				11031
				11201
				10100
				10102
				10103
				10104
EOF
if ( echo "${NODE_TYPE_LIST}" | grep CE > /dev/null ); then
cat <<EOF >>${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf
		TCP
			Server ${GRIDICE_SERVER_HOST}
			Port 12409
			FilterMetrics KeepOnly
				10106
				10107
EOF
fi
cat <<EOF >>${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf

	Sensors

		edtproc 
			CommandLine ${INSTALL_ROOT}/gridice/monitoring/bin/GLUEsensorLinuxProc
			MetricClasses
				edt.uptime  
				edt.cpu  
				edt.memory  
				edt.disk  
				edt.network  
				edt.ctxint  
				edt.swap  
				edt.processes  
				edt.sockets  
				edt.cpuinfo  
				edt.os  
				edt.alive  
				edt.regfiles  

		sensor1 
			CommandLine ${INSTALL_ROOT}/edg/libexec/edg-fmon-sensor-systemCheck
			MetricClasses
				executeScript  


	Metrics
		11001
			MetricClass edt.uptime 
		11011
			MetricClass edt.cpu 
		11021
			MetricClass edt.memory 
		11101
			MetricClass edt.disk 
		11202
			MetricClass edt.network 
			Parameters
				interface       eth0
		11013
			MetricClass edt.ctxint 
		11022
			MetricClass edt.swap 
		11031
			MetricClass edt.processes 
		11201
			MetricClass edt.sockets 
		10100
			MetricClass edt.cpuinfo 
		10102
			MetricClass edt.alive 
		10103
			MetricClass edt.regfiles 
		10104
			MetricClass executeScript 
			Parameters
				command ${INSTALL_ROOT}/gridice/monitoring/bin/CheckDaemon.pl --cfg ${INSTALL_ROOT}/gridice/monitoring/etc/gridice-role.cfg
EOF
if ( echo "${NODE_TYPE_LIST}" | grep CE > /dev/null ); then
if [ "X$GRIDICE_REFRESH_INFO_JOBS" = "Xno" ]; then
  OPT_REFRESH=" --no-update"
fi 
cat <<EOF >>${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf
		10106
			MetricClass executeScript
			Parameters
				command  ${INSTALL_ROOT}/gridice/monitoring/bin/CheckJobs.pl --lrms=${CE_BATCH_SYS} --lrms-path=${BATCH_LOG_DIR} --interval=${LAST_HOURS_EXEC_JOBS} --hide-subject=${GRIDICE_HIDE_USER_DN} --jm-dir=${JM_TMP_DIR} $OPT_REFRESH
EOF
cat <<EOF >>${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf
		10107
			MetricClass executeScript
			Parameters
				command ${INSTALL_ROOT}/gridice/monitoring/bin/LRMSinfo.pl --lrms=${CE_BATCH_SYS}
EOF
fi
cat <<EOF >>${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf

	Samples
		verylowfreq
			Timing 3600 0
			Metrics 
				10100
		lowfreq
			Timing 1800 0
			Metrics 
				11001
EOF
if ( echo "${NODE_TYPE_LIST}" | grep CE > /dev/null ) && [ "X$GRIDICE_JM" = "Xyes" ]; then
cat <<EOF >>${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf
		midfreq
			Timing 1200 0
			Metrics
				10106
EOF
fi
cat <<EOF >>${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf
		proc0
			Timing 30 0
			Metrics 
				10102
		proc1
			Timing 60 0
			Metrics 
				11011
				11021
				11101
				11202
				11022
				11031
				11201
		proc2
			Timing 300 0
			Metrics 
				10103
EOF
if ( echo "${NODE_TYPE_LIST}" | grep CE > /dev/null ); then
cat <<EOF >>${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf
				10107
EOF
fi
cat <<EOF >>${INSTALL_ROOT}/edg/var/etc/edg-fmon-agent.conf
		proc3
			Timing 120 0
			Metrics 
				10104
EOF

# Configure the job monitoring daemon only on CE 
if ( echo "${NODE_TYPE_LIST}" | grep CE > /dev/null ); then
  /sbin/chkconfig gridice_daemons on
  /sbin/service gridice_daemons stop
  /sbin/service gridice_daemons start
fi

/sbin/chkconfig edg-fmon-agent on
/sbin/service edg-fmon-agent stop
/sbin/service edg-fmon-agent start

# The cron job required was originally installed under
# the spurious name edg-fmon-knownhosts
if [ -f ${CRON_DIR}/edg-fmon-knownhosts ]; then
    rm -f ${CRON_DIR}/edg-fmon-knownhosts
fi

if [ "X$GRIDICE_SERVER_HOST" = "X`hostname -f`" ]; then
/sbin/chkconfig edg-fmon-server on
/sbin/chkconfig gridice-mds on
/sbin/service edg-fmon-server stop
/sbin/service edg-fmon-server start
/sbin/service gridice-mds stop
/sbin/service gridice-mds start

cron_job edg-fmon-cleanspool root "41 1 * * * ${INSTALL_ROOT}/edg/sbin/edg-fmon-cleanspool &> /dev/null"

#Clean up any remaining sensitive information
find /var/fmonServer/ -name 'last.00010101' -exec rm -f '{}' \;

fi


return 0
}

About this document ...

BDII Generic Configuration Reference

This document was generated using the LaTeX2HTML translator Version 2002 (1.62)

Copyright © 1993, 1994, 1995, 1996, Nikos Drakos, Computer Based Learning Unit, University of Leeds.
Copyright © 1997, 1998, 1999, Ross Moore, Mathematics Department, Macquarie University, Sydney.

The command line arguments were:
latex2html -split 0 -html_version 4.0 -no_navigation -address 'GRID deployment' BDII.drv_html

The translation was initiated by Oliver KEEBLE on 2006-01-16