ciips-network.html

   1 <HTML>
   2 <HEAD>
   3 <TITLE>Administrating the CIIPS Network - the easy way</TITLE>
   4 </HEAD>
   5 <BODY>
   6
   7 Petter Reinholdtsen &lt;pere@td.org.uit.no&gt;, 2000-04-17
   8
   9 <HR>
  10 <H1>Administrating the CIIPS Network - the easy way</H1>
  11
  12 <BLOCKQUOTE><EM>Some suggestions to make it easier to maintain and
  13 improve the CIIPS network.</EM></BLOCKQUOTE>
  14
  15 To maintain a multi-user network it is vital to reduse the possibility
  16 for human errors as much as possible.  It is also necessary to make
  17 sure any error get as little impact as possible.
  18
  19 <P>To achieve this, I believe the following must be done:
  20
  21 <DL>
  22
  23  <DT><STRONG>Keep things consistent</STRONG></DT>
  24
  25  <DD>Make sure knowledge is generic and not related to the different
  26  computers.  Make sure every supported program is available on all
  27  platforms and operating systems, with the same version and the same
  28  configuration. (<em>cfengine, Store</em>)</DD>
  29
  30  <DT><STRONG>Automate as much as possible.</STRONG></DT>
  31
  32  <DD>By leaving the repetitive tasks to the computer, one assures they
  33  are done the same way everywhere, every time. (<em>cfengine</em>)</DD>
  34
  35  <DT><STRONG>Make it possible to back out changes.</STRONG></DT>
  36
  37  <DD>When problems arise, make sure it is easy to get back to a
  38  earlier state when things where working.  Use version control systems
  39  to keep track of all human-edited configuration files, and make sure
  40  it is easy to remove a newly upgraded software package if problems
  41  are discovered. (<em>CVS, Store</em>)</DD>
  42
  43  <DT><STRONG>Detect problems as early as possible.</STRONG></DT>
  44  <DD>If potential problems are detected and fixed before they become
  45  problems, less work is required to fix the resulting domino-effect
  46  when one system fail.  (Example: a full disk can stop the backup
  47  system from working properly, or a failing NFS server can hang
  48  processes on other servers and eventually bring other servers to
  49  their knees).
  50
  51  <BR>Make sure to monitor all hosts and services, and warn when
  52  something is about to go wrong.  Keep statistics to find repeating
  53  problems. (<em>Palantir, mon</em>)</DD>
  54
  55  <DT><STRONG>Keep the users informed</STRONG></DT>
  56
  57  <DD>Make sure the users know where to find information on current
  58  problems, and the status of their reported problems.  This is easiest
  59  solved using a database and web based problem tracking
  60  system. (<em>Bugzilla</em>)</DD>
  61
  62  <DT><STRONG>Prepare for disaster</STRONG></DT>
  63
  64  <DD>Some suggestions:
  65  <UL>
  66
  67   <LI>Hard-disks fails after a few years.  Keep only new HDs in the
  68   servers, and reuse the old server HDs in workstations until they
  69   fail.
  70
  71   <LI>Back up everything, and back up often.  Make it easy to get
  72   files back from backup, and let individual users get their files
  73   back on their own.  It saves support personnel the hassle, and make
  74   it easier for the users to recover from their personal disasters.
  75
  76   <LI>Make it easy to replace workstations.  A HD mirroring system to
  77   make it trivial to reinstall a workstation and easy to replace a
  78   broken computer. (<em>Autosetup</em>)
  79
  80  </UL></DD>
  81 </DL>
  82
  83 <H2>Tools</H2>
  84
  85 <UL>
  86  <LI><A HREF="http://www.iu.hioslo.no/cfengine/">cfengine</A>
  87  <LI><A HREF="http://www.pvv.org/~arnej/store/storedoc.html">Store</A>
  88  <LI><A HREF="http://www.sourcegear.com/CVS/Dev/resource">CVS</A>
  89  <LI><A HREF="http://www.palantir.uio.no/cgi-bin/nodelist">palantir</A>
  90  <LI><A HREF="http://www.kernel.org/software/mon/">mon</A>
  91  <LI><A HREF="http://www.mozilla.org/bugs/">Bugzilla</A>
  92
  93  <LI>Autosetup is not publicly available, but I can get it from the
  94  author.  More expensive alternatives are also available.  <A
  95  HREF="http://jacal.sourceforge.net/summary.php">jacal</A> might be an
  96  alternative.
  97 </UL>
  98
  99 I've already installed Store in the robotics lab, and it will take ~15
 100 minutes to set it up on other hosts as well.  With proper setup, this
 101 will make most hosts more independent from the NFS servers.
 102 </BODY>
 103 </HTML>