SUMMARY 2: One Batch Queue feeding several Hosts

From: Jochen Bern (bern@kleopatra.Uni-Trier.DE)
Date: Mon Sep 06 1993 - 21:45:57 CDT

Yes - again. ;-)

Original Question:

> we are currently crunching a Number of heavy Jobs batch(1)ed by Shell
> Scripts. I investigated the Possibility of tuning Batch Queue Defs,
> but what I'ld really like to have is that the available Hosts (cur-
> rently five) would fetch their Jobs from a single Queue, instead of
> five Queues with manual Load Balancing.
> I forgot to mention that this is for SunOS 4.1.2, one SS2 and several ELCs.

Well, I tried DQS 2.1 - and didn't like it, for several Reasons:

1) Doesn't really feed multi-Host (!) because you have to designate one Queue
   as "Master Queue" of the Job.

2) The batch() Replacement has serious Problems parsing its Command Line.

3) Output is not caught and mailed. The only Thing you can do is to get it
   written to a File.

4) The "Job terminated" Mail is filled with some weird Control Char Litter.

5) My Test Jobs (#!/bin/csh with several echo's) returned only Error Messages,
   no Matter whether stdout/stderr was pointed to a File or not.

Unexpectedly, I received another Response from
(Adam Shostack) suggesting a Package called "Condor". I FTPed it down, but it
is somewhat different from what I wanted. It consists mainly of Replacements
for the Runtime Lib of C Programs, making it possible to parallelize your
Binaries - rather a Cousin of PVM, I think. Thanks, though.

Well, I'll try out NQS now. DQS referenced it in its Docs, so I hope that it's
a little more stable. Wish me Luck.

                                                                        J. Bern

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:08:09 CDT