If it is just for load balancing, I would go for a single LPAR.  
Easier to configure.
All resources are available where they are needed, when they are needed.

If you are looking for uptime, then you need at least two LPARs, so you could 
survive a LPAR outage.  And as you don't use VM, multiple LPARs would survive a 
Linux outage.

You can share CPU across LPARs.
You can share I/O across LPARs.
But you have to dedicate memory to LPARs.

Also, if you setup the IOCP to allow sharing of all you I/O devices across 
multiple LPARs, you will find that "control storage" requirements increases, 
which uses up more of your main memory.  This, of course, is not the case on 
the newer processors.

Also, multiple copies at a high level also allow for taking down parts for 
maintenance and/or reconfiguration and/or software upgrades and/or hardware 
upgrades, without a total outage.

Very quickly, this type of discussion (over beers), moves away from the techie 
considerations to the business considerations.

Tom Duerbusch
THD  Consulting


Law of Cat Sleeping

  All cats must sleep with people whenever possible, in a position
  as uncomfortable for the people involved, and as comfortable as
  possible for the cat


>>> Patrick Spinler <[EMAIL PROTECTED]> 12/7/2007 9:23 AM >>>
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


We're having a long running over a beer discussion about how best to
configure one of our applications.  During our last drinking session,
er, debate, we thought it might be interesting to post here and see
what, if any, consensus there is.

In brief: when load balancing client network connections across
multiple instances of an application, is it better to run multiple
separate 1 virtual IFL linux guests, each with a separate copy of the
application, or a single, large N virtual IFL linux guest with several
copies of the application running inside it?

Assumption: you have enough real IFL's to back the number of virtual
IFL's you'd assign.

Our real situation which gave rise to this question:  We're running
the tivoli monitoring suite of products on our z system in linux
guests.  As per IBM's documented recommendation, we're running one
TEMS process for each apx 500 client systems we monitor.  Currently,
we have this configured as a number of separate linux guests, each
running it's own copy of TEMS.

With expansion we have budgeted for next year, though, we'll have
enough real IFL's to consider consolidating all these TEMS proceses
into a single, multi-virtual-ifl linux guess.

We're not sure if there would be any cp scheduling benefits to either
configuration, or any resource benefits (e.g. storage usage, or paging).

Help support our fluid fueled debates !  What're the group's thoughts,
please?

Thanks!
- -- Pat
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org 

iD8DBQFHWWWANObCqA8uBswRApO7AJ40MdZJtQnI0enpwLxV+FLAjcrG2ACfZwOC
xaoLzQy/+vMMFNWgtETqLrI=
=0gIM
-----END PGP SIGNATURE-----

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to