One of the things you have to keep in mind with PVM is that it uses two
files in /tmp to keep track of its resources. pvmd.X and pvml.X - where
X is the UID of the user.

Either one or two things are happening here, and this depends on whether
or not your workstation's /tmp (or rather, what pvm THINKS your
workstation's /tmp) is shared across multiple machines. 

If PVM is running or happens to die or is inappropriately stopped (using
anything else but halting pvm manually), these files will be left behind
and the userland daemon will fail to start during subsequent attempts.
In this case, all you have to do is to get rid of the user's pvmd file
on that machine, and you should be fine. You can probably do this
through a shell script that the user can either call manually and / or
at workstation start up (eg. when s/he logs in). However you deal with
this, DON'T do a blanket rm -rf pvmd.* unless you want to really want to
irritate other users on the cluster. I have a shell script here you
might be able to use after some modifications, let me know if you want it.

If your workstations use the same /tmp as the server, the server
attempting to start on the workstation will notice the pvmd file, and
stop, as you see. You basically will have to set up either a local
/tmp on each workstation or nfs mount a SEPARATE partition for each
machine. Pain in the arse, yes, but that resource doesn't have to be
very large.

Hope this helps. PVM is cool, but it certainly does have some
interesting quirks

-- pete

On Fri, Nov 01, 2002 at 05:52:23AM -0800, Athar Nouman wrote:
> 
> Now I am trying to run the pvm on my server and trying to add the host (diskless 
>workstation). But the pvm does not search the deomon on the workstation instead it 
>loops back to the server and generats a message as follows:

> Determining PVM Temporary Directory on Host "ws001"
> 
> $PVM_TMP on ws001 set to /tmp
> 
> Checking for Leftover PVM Daemon Files on Host "ws001"
> 
> PVM Daemon Files Found on ws001!
> 
> Either PVM is Already Running or else it crashed and left behind a /pvmd.<uid> 
>daemon file.
> 
> Halt PVM if it is running on ws001, or else remove any leftover /pvmd.* files.

-- 
Pete St. Onge
Research Associate, Computational Biologist, UNIX Admin
Banting and Best Institute of Medical Research
Program in Bioinformatics and Proteomics
University of Toronto
http://www.utoronto.ca/emililab/


-------------------------------------------------------
This sf.net email is sponsored by: See the NEW Palm 
Tungsten T handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0001en
_____________________________________________________________________
Ltsp-discuss mailing list.   To un-subscribe, or change prefs, goto:
      https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help,   try #ltsp channel on irc.openprojects.net

Reply via email to