1. 05 Dec, 2006 1 commit
  2. 01 Dec, 2006 1 commit
  3. 09 Nov, 2006 1 commit
    • Mike Hibler's avatar
      Largely synchronize with the template version: · a94f74e1
      Mike Hibler authored
       * allow for no server program to be run
       * allow for a program to be run on all nodes (including planet-*)
       * mods for the new improved dbmonitor
      
      Also allow setting of use_magent=2 to indicate that it should start an
      iperfd to listen for traffic on plab nodes rather than the stub.
      a94f74e1
  4. 09 Oct, 2006 1 commit
  5. 03 Oct, 2006 1 commit
    • Mike Hibler's avatar
      First crack at moving toward a template "run" model. · 8f87a63b
      Mike Hibler authored
      Create event sequences to do the equiv. of start- and stop-experiment.
      Rather than using those scripts, you now can use start-run and stop-run
      which are just wrappers around tevc to trigger the sequences.
      
      This checking is backward compatible in the sense that you can continue to
      use start- and stop-experiment and should never notice a thing.
      
      The new event model doesn't support a couple of things that the old script
      based execution did, most notably, you cannot change command line arguments.
      It does support real vs. fake runs.
      8f87a63b
  6. 29 Sep, 2006 2 commits
    • Robert Ricci's avatar
      Allow the setting of ports for the stub, so that two experiments can · 6ad24846
      Robert Ricci authored
      run simultaneously on the same planetlab node.
      
      Confine tcpdump on the planetlab nodes to just the stub's traffic, so
      we don't have to worry about nameserver traffic, ssh traffic, etc.
      
      Make use_magent the defauly.
      6ad24846
    • Mike Hibler's avatar
      Add no_plab option to NS file allowing you to "monitor" (set initial · a7434c80
      Mike Hibler authored
      conditions) plab nodes without actually allocating them.
      
      Changed init-elabnodes to support this.  Elab to plab mapping is set in the
      NS file via the $opt array, which init-elabnodes extracts via the XMLRPC
      interface.
      
      So to run the "really simple model":
      
       1. set "no_plab" to 1 in the NS file
       2. make sure you have enough nodes listed in "plabnodes" in the NS file
          to cover "num_pcs".
       3. swapin the experiment
       4. after swapin, run init-elabnodes.pl.
      
      Also, cleaned up the NS file some: got rid of hokey tcpdump setup as it
      was redundant, made tracing optional (though on by default), expect proper
      return values from program agents.
      a7434c80
  7. 14 Sep, 2006 1 commit
  8. 01 Sep, 2006 1 commit
  9. 25 Aug, 2006 1 commit
  10. 16 Aug, 2006 1 commit
  11. 12 Aug, 2006 1 commit
  12. 11 Aug, 2006 2 commits
  13. 10 Aug, 2006 1 commit
    • Mike Hibler's avatar
      First crack at surviving down planetlab nodes. If the master barrier sync · 5f413b47
      Mike Hibler authored
      node sits in the stub or monitor barrier sync for more than the SYNCTIMO
      timeout value in common-env.sh, it will send a HUP to syncd which will
      knock all the other nodes out of their barrier sync.  If that happens,
      all nodes will print a warning message and continue.
      
      All nodes wait for both a stub sync and a monitor sync, so if one plab node
      is down, they will timeout on both barrier syncs.  Race conditions?  Sure.
      If for example everyone times out on the stub barrier due to a slow node,
      and then that node reaches the barrier, it will hang there while everyone
      else waits on the monitor barrier.  When the latter times out, it will
      kick the slow node out of the stub sync and it will then proceed to hang
      in the monitor sync until the experiment is stopped.  Got that?
      
      As an aside, it would be nice if the initializer of a barrier could specify
      a timeout value, and return a special error code to everyone if it timed out,
      but that would require an incompatible change to the sync protocol.
      5f413b47
  14. 09 Aug, 2006 3 commits
  15. 19 Jul, 2006 1 commit
  16. 25 Apr, 2006 2 commits
  17. 23 Apr, 2006 1 commit
  18. 20 Apr, 2006 1 commit
    • Mike Hibler's avatar
      Various changes: · 73d52a50
      Mike Hibler authored
       * auto-pelab.ns: add an event-group that runs 'update -t' on the nodes
         to update the tarballs/RPMs.  NOTE that you will need to modify your
         experiment and update your version of the NS file for this to work.
      
       * install-tarball: simple script to run tevc and trigger the tarball
         updates
      
       * start-experiment: '-p' option to monitor the real planetlab nodes
         instead of the emulated ones, -M/-S to set the arguments passed to
         the monitors/stubs, -m/-s to append to the current arguments passed
         to the monitors/stubs.
      
       * monitor/auto-monitor.sh: pass through the command line arguments
      73d52a50
  19. 18 Apr, 2006 1 commit
  20. 14 Apr, 2006 2 commits
  21. 12 Apr, 2006 2 commits
  22. 10 Apr, 2006 1 commit
  23. 05 Apr, 2006 1 commit
    • Robert Ricci's avatar
      Framework for running pelab experiments. · d174b055
      Robert Ricci authored
      CONTENTS:
      
      auto-pelab.ns - NS script for creating a pelab experiment. Supports
      both real and fake Plab nodes.
      
      start-experiment - perl script to start up monitors and stubs, and to
      start link tracing.
      
      stop-experiment - perl script to stop monitors and stubs, and collect
      all of the files from this experiment.
      
      TODO:
      Write documentation for it - how to run it, where the logfiles go, etc.
      
      Add hooks for the user to run their own program under test (ie iperf)
      
      Reset condidtions on the Emulab-side LAN before an experiment starts
      
      Get link tracing on plab (probably by starting pcapper with the
      program agent)
      d174b055