1. 26 Nov, 2003 2 commits
  2. 25 Nov, 2003 5 commits
  3. 24 Nov, 2003 8 commits
  4. 23 Nov, 2003 1 commit
  5. 21 Nov, 2003 4 commits
    • Mike Hibler's avatar
      Tweaks to support e1000 cards on the pc2000s. Installing this will break · 21e9611f
      Mike Hibler authored
      pc170-1 until those nodes actually have e1000 cards installed (next week).
      21e9611f
    • Mike Hibler's avatar
    • Mike Hibler's avatar
      Make sure we can hardwire speed/duplex on new Intel Pro/1000 (e1000) cards. · 5ad5ab99
      Mike Hibler authored
      Need to use ethtool instead of mii-tool to do this.  My comment:
      
        #
        # Linux is apparently changing from mii-tool to ethtool but some drivers
        # don't support the new interface (3c59x), some don't support the old
        # interface (e1000), and some (eepro100) support the new interface just
        # enough that they can report success but not actually do anything.  Sweet!
        #
      
      This requires that we load ethfind on the nodes, though the script will
      continue to work if it isn't (though will not work for e1000 cards).
      5ad5ab99
    • Mike Hibler's avatar
      Hack fix for an apparent boot-time race condition. Our pump-invoked · 633a7a12
      Mike Hibler authored
      sethostname script properly sets the hostname, but sometimes that hostname
      gets overridden by one of the static boot scripts which sets the hostname
      back to "localhost".  The sequence is something like:
      	pump runs on eth0 and fires off our sethostname script
      	sethostname blocks
      	ifup of eth1 starts, determines that hostname hasn't been
      	  set (i.e., is set to "localhost") and remembers (NEEDHOSTNAME)
      	sethostname finishes by setting the host name
      	ifup of eth1 finishes, seeing NEEDHOSTNAME is set but doesn't
      	  have anything to set it to so resets to "localhost"
      Weird.  Hack is to make sure it never thinks it needs to set the hostname
      by setting it to something that is not "localhost".  We do this once we
      identify the control net interface (and thus know we will be running pump RSN).
      633a7a12
  6. 20 Nov, 2003 4 commits
  7. 19 Nov, 2003 3 commits
  8. 18 Nov, 2003 12 commits
  9. 17 Nov, 2003 1 commit
    • Leigh B. Stoller's avatar
      Merge the two state machines (batchstate and state) into a single · 2025e0bd
      Leigh B. Stoller authored
      state machine (state). All of the stuff that was previously handled by
      using batchstate is now embedded into the one state machine. Of
      course, these mostly overlapped, so its not that much of a change,
      except that we also redid the machine, adding more states (for
      example, modify phases are now explicit. To get a picture of the
      actual state machine, on boss:
      
      		stategraph -o newstates EXPTSTATE
      		gv newstates.ps
      
      Things to note:
      
      * The "batchstate" slot of the experiments table is now used solely to
        provide a lock for batch daemon. A secondary change will be to
        change the slot name to something more appropriate, but it can
        happen anytime after this new stuff is installed.
      
      * I have left expt_locked for now, but another later change will be to remove
        expt_locked, and change it to active_busy or some such new state name in
        the state machine. I have removed most uses of expt_locked, except those
        that were necessary until there is a new state to replace it.
      
      * These new changes are an implementation of the new state machine,
        but I have not done anything fancy. Most of the code is the same as
        it was before.
      
      * I suspect that there are races with the batch daemon now, but they
        are going to be rare, and the end result is probably that a
        cancelation is delayed a little bit.
      2025e0bd