[LON-CAPA-admin] lon-balancer

[Lucas, Mark]

Hi,

I guess I’ve been having better luck with the balancer.

I’ve got capa5 (ohioua1) set up as a dedicated balancer.

ohioua2, ohioua3, and ohioua4 are all primary (you can have multiple primary).

ohioua4 is listed as default.

These are all dedicated, single machine access servers.

I offload advanced users and ohiou authors to oucapa2 (our library server).

I have a web of “user session offloading” where various machines offload to various
other machines when busy. If the balancer is balancing, this typically isn’t an issue.

I then set ohioua2, ohioua3, and ohioua6 to redirect to capa5 (ohioua1, the balancer)
if there is a login attempt.

My monitoring of active users (I graph the number of users active in the last 15 minutes using
mrtg and a quick script) show a pretty good balance between the 3 primary
access servers.

Could it be that people are not defining several primary machines from their balancer?
My understanding is that the default only gets called when the load gets too high on
the primary machines.

Later,
Mark


On Oct 7, 2015, at 4:28 PM, Todd Ruskell <todd.ruskell at gmail.com> wrote:

> Hi All,
> 
> Unfortunately, I don't have any suggestions, because I've been noticing the same thing:
> 
> My domain has three servers--a lonbalancer, a library server, and an access server.  My intent is to maintain roughly equal loads on the library and access servers, both of which are VMs with 8 CPUs and 16 GB of ram.  Both have a lonLoadLim of 5.
> 
> I know that the balancer can and does offload sessions to both the library and access servers, both of which are designated as "default" servers.  But it doesn't seem to do so in a way I expect.  My understanding is that new sessions are supposed to go the server with the least load.  Yet, with conditions similar to these:
> 
> library server load average:  0.38, 0.56, 0.53
> userstatus.pl reports: 28 Active and 214 Moderately active users
> 
> access server load average: 0.01, 0.04, 0.01
> userstatus.pl reports: 6 Active and 30 Moderately active users
> 
> user sessions seem to always go to the library server.  It is not until the 1-minute load average drops to zero that the load balancer directs a session to the access server.  I seem to have the same behavior when both servers are set to "primary" as well.  It doesn't seem like this is how it's supposed to work.  
> 
> What am I missing?
> 
> Any help you can provide would be greatly appreciated.
> 
> Thanks,
> Todd
> 
> On Wed, Oct 7, 2015 at 12:57 PM, Abdel Messeh, Maged <mmesseh at illinois.edu> wrote:
> Hi All,
> 
> I have recently noticed that we have 4 or 5 times more requests going to one of our access nodes than the other 3.
> 
> Is there a way of checking if the balancer working right? And fixing it, if not?
> 
> Many thanks,
> 
> Maged
> 
> 
> --------------------------
> Maged Messeh, Ph.D.
> College of Liberal Arts & Sciences Infrastructure
> University of Illinois at Urbana-Champaign
> 
> 
> _______________________________________________
> LON-CAPA-admin mailing list
> LON-CAPA-admin at mail.lon-capa.org
> http://mail.lon-capa.org/mailman/listinfo/lon-capa-admin
> 
> _______________________________________________
> LON-CAPA-admin mailing list
> LON-CAPA-admin at mail.lon-capa.org
> http://mail.lon-capa.org/mailman/listinfo/lon-capa-admin

-- 
Mark Lucas 								email: lucasm at ohiou.edu
252D Clippinger Lab						phone: (740)597-2984
Department of Physics and Astronomy			fax: (740)593-0433
Ohio University
Athens, OH 45701