Remstats Home

  About remstats
  Release Notes
      the remstats user
      the webgroup group
      private functions
      remstats servers
      your web-server
  Configuration Tools

Live Data

Remstats was written by Thomas Erskine at the CRC in Canada and now looking for work.

[Top] [Prev] [Next]

How to install remstats

READ THE RELEASE NOTES FIRST. This page is generic and does NOT include version-specific instructions.

I know that this is not simple. I do plan to make it simpler, but it'll never be "./configure; make; make install" because I don't know what you want to monitor.

The two C programs (multiping and traceroute) now use autoconf, and the main configure script works (from the outside) simlarly to an autoconf-generated configure. I haven't seen a need to convert it to autoconf yet. It's mostly perl scripts and if you have the right version of perl properly installed, it shouldn't need anything special. The unix-status-server is a slight exception to this, but the only configuration needed so far is done dynamically and is only the location of the various required utilities.

  1. Unpack the distribution tarball:
    	gunzip -dc remstats.tar.gz | tar xf -
  2. Create the remstats user and group, if you haven't already, (by default remstats and remstats respectively.) (See also the remstats user.) Unless you plan to run a separate instance of the webserver for remstats, you'll also want to create the webgroup group.

  3. Build and install the software. If you're upgrading, you might want to take a copy of fixup.config from the old version:
    	sh configure

    If you want to override the defaults, then run

    	sh configure --help

    for a list of what can be overridden.

    [Check fixup.config to make sure it is properly setup.]

    	make all
    	make upgrade
    	make install
    	su - 'make install-suid'

    Note: this step also customizes the programs and documentation with your choice of directories, owner, ... so this documentation should refer to your setup after you've done the install.

    The make install-suid simply makes traceroute and multiping suid root. They won't work most places unless run as root, one way or another. Since I don't like to run all of remstats as root, this was the best compromise I could come up with.

  4. Fix the config-base for site-specific things. Edit the following files in /home/groups/r/re/remstats/etc/config-base, looking the the string "FIXME", without the "quotes".
    	alerts general html scripts/http-proxy

    I'll try to keep this list up to date, but you can make sure by doing:

    	grep -l FIXME /home/groups/r/re/remstats/etc/config-base/* /home/groups/r/re/remstats/etc/config-base/*/*
  5. Make a config-dir to describe what you want to monitor. You can do this by hand, or using the configuration building tools. To use the tools, you'll have to make a few files listing various kinds of hosts:
    	cd /home/groups/r/re/remstats/etc
    	/home/groups/r/re/remstats/bin/new-config config
    	/home/groups/r/re/remstats/bin/new-ping-hosts groupname1 group1-hosts-file
    	/home/groups/r/re/remstats/bin/new-ping-hosts groupname2 group2-hosts-file
    	/home/groups/r/re/remstats/bin/new-port-hosts groupname3 port-hosts-file
    	/home/groups/r/re/remstats/bin/new-snmp-hosts groupname4 SNMP-community-string snmp-hosts-file

    After you've installed the unix-status-server on some hosts, you can also use:

    	/home/groups/r/re/remstats/bin/new-unix-hosts groupname5 unix-hosts-file

    If you have any Windows NT hosts that you want to monitor, after you have installed the nt-status-server, you can run nt-discover to find and add the NT hosts for a given NT domain.

    If you're going to use the log-collector, you'll have to build the rrd entries for each by hand. There doesn't seem to be much standard in where log-files go, let alone what's in them.

  6. Arrange for cron to run run-remstats2 at an appropriate interval. For a five-minute interval, something like the following will do:
    	0,5,10,15,20,25,30,35,40,45,50,55 * * * * /home/groups/r/re/remstats/bin/run-remstats2

    This checks the configuration, collects the new data, updates the rrds, runs the monitors to compute statuses and updates the web-pages. Note: it does not re-write the web-pages for every iteration; it only does so when configuration files have changed, as the web-pages will show new data by themselves. This and more are in the sample file under /home/groups/r/re/remstats/etc.

  7. [optional] Arrange for cron to run do-traceroutes at an appropriate interval. You could run it in the wee hours of each morning like:
    	5 3 * * * /home/groups/r/re/remstats/bin/do-traceroutes

    This information isn't currently used, but I'm planning to make use of it.

  8. [optional] Arrange for cron to run snmpif-description-updater periodically, if you have any snmpif-* RRDs, which you're likely to change the descriptions on. Say every day, like:
    	0 3 * * * /home/groups/r/re/remstats/bin/snmpif-description-updater
  9. Arrange for cron to run remstats-cleanup every now and then to remove old un-needed files, like:
    	0 3 * * * /home/groups/r/re/remstats/bin/remstats-cleanup

    This removes no-longer-needed files, like old host graphs, traceroute results, log-files, ... It does not remove rrd and status files, on the principle that you might want them and they can't be re-created.

  10. You'll need to set up your web-server to allow CGI scripts in the remstats html tree and make sure that you're not allowing everyone in.

  11. Make a symlink in the html directory from whichever index you prefer to index.cgi.

  12. You'll want to look at the server installation docs if you're going to be running any of the remote servers ( log-server, nt-status-server, remoteping-server, and unix-status-server).

Enjoy your pretty pictures and I hope that you find them usefull.

[Top] [Remstats] [SourceWorks] [RRDtool] [SourceForge]
Last updated Fri May 30 13:50:56 PDT 2003 by <>.