Re: [Hive-users] re: CellListAgent

[Raffi K. <rc...@mi...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


hi!  the problem is that the main server list agent is not running on
pinotnoir.media.mit.edu...

there are three solutions to the problem

1) run the master server list agent. i'll start doing that again.
   i'll try to make sure that hive 19991210 is always running on
   pinotnoir.medis.mit.edu with the master server list agent

2) you could run your own master server list agent.  you can designate
   this at the command line by doing a

   java edu.mit.media.hive.Hive --join hive://sc258.kings.cam.ac.uk:6363 

   or whatever your machine is.  this tells your server list agent to
   join a federation managed by itself.  and if you start up other
   hive cells and point it to that cell, then the server list agent on
   that cell will act as the master for everythign

3) tell your server list agent to not join any federation.  you can do
   that by running
  
   java edu.mit.media.hive.Hive --nojoin


of all these, the ideal solution is #1.  and once i get into lab today
(in a few hours), i'll get the master server list agent running again.

feel free to e-mail out again if you need anything else!

rk



Suranga Chandratillake writes:
 > 
 > Hi all,
 > 
 > I've been playing with Hive for a couple of months now and never had any
 > real problems, but recently I get a hang while a new cell is being
 > setup.  Specifically, it happens when CellListAgent issues the following
 > message :
 > 
 > <NOTICE CellListAgentImpl.java:441> Tracking
 > hive://sc258.kings.cam.ac.uk:6363 2Apr2000-04:04:38.030
 > 
 > Basically, the code hangs at this point for a variable length of time (the
 > worst ever was about five minutes) and then proceeds as normal.  I'm not
 > sure, but I get the feeling this is to do with the code looking at the
 > default CellList url
 > (nelson.www.media.mit.edu/people/nelson/tmp/hiveMasterCell) and that
 > taking a long time to come through.  The main reasoning behind this is
 > that if I try and look at Nelson's pages on my web browser, they seem to
 > take a similar time too.
 > 
 > Anyway, does anyone know if this diagnosis is correct and, if so, what
 > does the hiveMasterCell file actually look like and could I just re-create
 > it on my machine and point the code towards that instead while I'm
 > coding/testing?
 > 
 > Thanks,
 > 
 > Suranga

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.1 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE46JBsEbM0gNEaRkURAvKjAKCGY801jKjTBFw7GJRceQJwQbkeZwCeLsuy
KD39neIeiMVJxEuhzHvjl3g=
=GCkA
-----END PGP SIGNATURE-----