Scalable PBS

Chris Samuel csamuel at
Thu Aug 14 21:52:06 EDT 2003

Hash: SHA1

Hi all,

Just joined the list, so apologies if this is already well known.

I noticed a recent message in the archive about OpenPBS and problems with 
scalability, and I think it's worth noting that there is an alternative (and 
actively developed) fork of OpenPBS called "Scalable PBS" available from:

Amongst other features it has (quoting the website):

	Better Scalability
	   - Significantly improved server to MOM communication model, the ability to
	     handle larger clusters, larger jobs, larger messages, etc.
	   - Scales up to 2K nodes vs ~300 nodes for standard OpenPBS.
	Improved Usability by incorporating more extensive logging, as well as, more
	human readable logging(ie no more 'error 15038 on command 42').

We're using SPBS here at VPAC on our IBM cluster and it's a lot better than 
the last OpenPBS release (2.3.16, from 2001). They forked off from 2.3.12 
rather than the last OpenPBS because it had a more open license.

The folks behind the project have worked very quickly with us to fix bugs 
we've been finding in it, typically when I found a bug they had fixed it 
within a day or so, usually overnight from my perspective in Oz. :-)

If you are considering using it I'd suggest using the current snapshot release 

as that irons out a couple of bugs that might bite.

For the less adventurous there is a new release SOpenPBS-2.3.12p4 due out in 
the near future that will include the fixes from the current snapshot.

- -- 
 Chris Samuel -- VPAC Systems & Network Admin
 Victorian Partnership for Advanced Computing
 Bldg 91, 110 Victoria Street, Carlton South,
 VIC 3053, Australia -

Version: GnuPG v1.2.2 (GNU/Linux)


Beowulf mailing list, Beowulf at
To change your subscription (digest mode or unsubscribe) visit

More information about the Beowulf mailing list