Seting up EPICS logging using the ChannelArchiver

The data logged by the viewed either by the X-Window application Xarr or over the web with the CGI program CGIExport. Either of these programs can also be used to extract text formated data useful spreadsheets and other applications.

Below we detail the steps followed to compile and configure the ChannelArchiver and viewer applications for the G0 target. These details are given both as a form of collaboratino documentation and in the hopes that these details can help setup EPICS logging for other applications.

The applications were installed on the account g0targ on the machine gzerol2.acc.jlab.org. This machine is a PC running RedHat Linux version 6.2.

Installing and Compiling Software

1. Add the followign to ~/.bash_profile

   export HOST_ARCH=Linux 
   export XARR_DB_INDEX=/home/g0targ/Archives/ArchiveData/freq_directory
   

   (Or do equivalent if using tcsh instead of bash)
2. Obtain car-xarr-cgiexport.tgz from Stephen Wood. From the home directory untar this file which will create the ~/car-xarr subdirectory.
3. Unpack Epics Version R3.13.0.beta12CEBAF3 into ~/epics. A tar file from Stephen Wood can be obtained for this.
4. Edit ~/epics/R3.13.0.beta12CEBAF3/config/RELEASE. Set EPICS_BASE and TEMPLATE_TOP:

   EPICS_BASE=/home/g0targ/epics/R3.13.0.beta12CEBAF3/base
   TEMPLATE_TOP=/home/g0targ/epics/R3.13.0.beta12CEBAF3/base/templates/makeBaseApp/top
   

   (Replace /home/g0targ if different)
5. Compile epics

   cd ~/epics/R33.13.0.beta12CEBAF3/base
   make
   

6. Get lesstif-0.89.9-2.i386.rpm and lesstif-2.0-devel-0.89.9-2.i386.rpm from /u/mirror/redhat/powertools/6.2/i386/i386 (Any CUE machine). Get gnuplot-3.7.1-3.i386.rpm from /u/mirror/redhat/redhat-6.2/i386/RedHat/RPMS. Install these from root account.
7. Build car and xarr with

   cd ~/car-xarr
   make
   

8. Copy all the relevant binaries to ~/bin or somewhere else in your PATH.

   cp -p ~/car-xarr/CAR/bin/Linux/ArchiveEngine ~/bin
   cp -p ~/car-xarr/XARR/bin/Linux/xarr ~/bin
   

Configuring and Starting the Archiver

1. Create the directories for the arhiver configuration and data

   mkdir ~/Archives
   mkdir ~/Archives/ArchiveData
   mkdir ~/Archives/cfg
   

   Note, the ArchiveData directory is going to hold all the output of the logger. It will get large, so it might be a good idea that ArchiveData be a softlink to a directory on a really huge disk.
2. Create scripts to start and stop the archiver. Make sure that the script to start the archiver sets the following environment variables: EPICS_TS_MIN_WEST EPICS_CA_AUTO_ADDR_LIST EPICS_CA_ADDR_LIST Here is an example ~/Archives/start_archiver script:

   #!/bin/bash
   
   export EPICS_TS_MIN_WEST=300
   export EPICS_CA_AUTO_ADDR_LIST=YES
   export EPICS_CA_ADDR_LIST="129.57.211.255 129.57.215.255 129.57.239.255 129.57.243.255"
   
   ARCHIVE_ENGINE_NAME=ArchiveEngine
   ARCHIVE_ENGINE=/home/g0targ/bin/$ARCHIVE_ENGINE_NAME
   ARCHIVE_DIR=/home/g0targ/Archives
   
   cd $ARCHIVE_DIR
   
   alreadyrunning="`ps auxww | grep $ARCHIVE_ENGINE_NAME | grep -v grep`"
   if [ "$alreadyrunning" = "" ] ; then
       $ARCHIVE_ENGINE -log $ARCHIVE_DIR/archive.log ./config $ARCHIVE_DIR/ArchiveData/freq_directory > archive.out 2>&1 &
   else
       echo "The Archiver is already running"
   fi
   

   Here is an example ~/Archives/stop_archiver script:

   #!/bin/bash
   
   ARCHIVE_ENGINE_NAME=ArchiveEngine
   ARCHIVE_ENGINE=/home/g0targ/bin/$ARCHIVE_ENGINE_NAME
   ARCHIVE_DIR=/home/g0targ/Archives
   ARCHIVE_LOCK_FILE=$ARCHIVE_DIR/archive_active.lck
   
   killall $ARCHIVE_ENGINE_NAME
   
   sleep 1
   
   stillrunning="`ps auxww | grep $ARCHIVE_ENGINE_NAME | grep -v grep`"
   if [ "$stillrunning" != "" ] ; then
       killall -9 $ARCHIVE_ENGINE_NAME
   fi
   
   if [ -f $ARCHIVE_LOCK_FILE ]; then
       rm -f $ARCHIVE_LOCK_FILE
   fi
   

3. Setup the configuration files "~/Archives/config" and any files it references. As an example, ~/Archives/config:

   # Group: config
   # This file was auto-created by the ChannelArchiver Engine
   # Instance: EPICS Channel Archiver Engine
   
   !write_period   30
   !default_period 1
   !get_threshold  300
   !file_size      24
   !buffer_reserve 3
   
   !group demo
   

   and ~/Archives/demo

   # Group: demo
   
   hallc:bcm1 1 monitor
   hallc:bcm2 1 monitor
   

4. Start the archiver running by executing your start_archiver script Make sure it is running by looking for 3 instances of "ArchiveEngine" running.

Setting up XARR

Type xarr.

Setting up CGIExport

1. Type the following commands:

   mkdir ~/public_html
   mkdir ~/public_html/cgi
   mkdir ~/public_html/cgi/tmp
   chmod 777 ~/public_html/cgi/tmp
   chmod 755 ~
   cp ~/car-xarr/CAR/bin/Linux/CGIExport ~/public_html/cgi/CGIExport.cgi
   cp -p ~/car-xarr/CAR/CGIExport/src/Tests * ~/public_html
   

2. Install the apache web server. Apache may already be installed on the system. Edit /etc/httpd/conf/httpd.conf, removing the "#" on the line "#AddHandler cgi-script .cgi".

   After "UserDir public_html" in httpd.conf, add:

   
           AllowOverride AuthConfig Limit
           Options Indexes ExecCGI
   
   

3. Restart web server with /etc/rc.d/init.d/httpd restart
4. In ~/public_html, create an index.html file with a link to CGIExport. The following index.html is a simple example:

   <html>
   <body>
   <a href="cgi/CGIExport.cgi?DIRECTORY=..%2FArchives%2FArchiveData%2Ffreq_directory&COMMAND=START">Archiver</a>
   </body>
   <html>
   

5. Customize cgi_body_start.txt and cgi_body_end.txt in ~/public_html. In particular, it is a good idea to add a comment to cgi_body_end.txt indicating who maintains this particular installation of CGIExport. For example, after the line with the author's (Kay-Uwe Kasemir) email address, add:

   %lt;br%gt;Installed for G0 by
   <a href="mailto:saw@jlab.org">Stephen A. Wood%lt;/a%gt;
   

6. Visit http://gzerol2.acc.jlab.org/~g0targ/ and play with the CGIExport utility.
7. Each time a plot is make with CGIExport, files are left behind in ~/public_html/cgi/tmp. These files need to be periodically deleted. A cron job could be used to do this automatically every day.
8. Access to the CGIExport display tool can be limited by the normal .htaccess mechanisms of apache. For example, the following .htaccess file limits access to users from machines explicitely listed on "allow from" lines and to those who know a username/password defined in the .htpasswd file. .htpasswd is maintained with the htpasswd command.

   <Limit GET>
   order deny,allow
   deny from all
   allow from gzerol2.acc.jlab.org wood.jlab.org
   allow from gzero.jlab.org
   AuthType Basic
   AuthUserFile /home/g0targ/public_html/.htpasswd
   AuthGroupFile /dev/null
   AuthName ByPassword
   require valid-user
   satisfy any
   </Limit>
   

Other Documentation

Xarr manual