Thanks for the responses (I'll take more!)
The options provided so far include:
- Portable Batch System
- Generic NQS
- Aurema (built into v5.1A to be release soon)
I'm not sure if it's a "must-have" but one of the users would like to have
NT support also (like LSF), so that jobs could be monitored from NT and/or
submitted to idle NT processors.
Here's a summary of the comments received:
................................................
Check out Portable Patch System (PBS) =
http://pbs.mrj.com/
It is a good, and open source, alternative to LSF.
................................................
Take a look at
http://www.aurema.com/ their software is being incorporated
into Tru64...it might just be the thing you need.
................
5.1A, which is ending FQ next week, has the Aurema system built in.
5.1a probably will ship in June.
The basic functionality is bundled and the hooks integrated into Tru64,
other functionality is 3rd party add-on.
aka ARMTech and Server Resource Management (don't know what it will be
called when it hits the street.
..................
There is Generic NQS, an open source network queueing system:
www.gnqs.org
It has a lot of functionality, including scheduling, but you don't have to
use all of its functionality: you can just set up one or more simple batch
queues. I don't know if it will make use of any cluster functionality.
................................................
Rob Aldridge
AT&T Solutions
Alliance, Ohio
-----Original Message-----
From: Aldridge, Robert E. [mailto:REAldridge_at_mcdermott.com]
Sent: Thursday, May 17, 2001 3:57 PM
To: tru64-unix-managers_at_ornl.gov
Subject: Batch scheduling / load-sharing
Importance: Low
Tru64 Managers:
I have an engineering/high-performance computing environment. Users like to
submit lots of analysis jobs (~ 100) and have them all get executed, be able
to monitor them, kill off them, etc.
First -- the engineers/users know about Platform Computing's Load Sharing
Facility (LSF) and would like us to evaluate that product. Do you have any
comments about using LSF on Tru64/TruCluster?
Second -- are there other built-in, freeware, or commercial packages that
provide for this type of batch queueing --- ESPECIALLY any packages that
take advantage of the TRUCLUSTER infrastructure ?
We're NOT looking for a complex scheduling package (e.g. we don't need to
run lots of interacting/interdependent overnight processes).
Thanks for your assistance/input/advise!
Environment: Tru64 5.1 patch 3; TruCluster 5.1; Dual ES40 systems via Memory
Channel
Rob Aldridge
AT&T Solutions
Alliance, Ohio
Received on Thu May 17 2001 - 21:17:29 NZST