argus max connection

[CS Lee]

hi Carter,

I can't run radium on mgt0 because mgt0 interface is only 1GbE and the data
sent from argus to radium is more than 1Gbps. My plan was having radium on
bivio connects to multiple argi and act as central point of access to my
analytic server. I have another analytic server that is running all the
argus client which is supposed to connect to radium, By the way I can't use
single core of radium to connect to argus, it will shoot up to 100% for
sure since my monitoring environment is hitting 10G.

Initially I was running argus and radium on same bivio box, however all 12
cores in bivio is hitting 100% and bivio crashed, I want to offload the
radium to another bivio box instead because after some testing, if I run
argus only on bivio, the CPU usage stays below 70% all the time.




On Fri, Jul 19, 2013 at 8:08 PM, Carter Bullard <carter at qosient.com> wrote:

> Hey CS Lee,
> Well that is curious.  Is it just because they all share the same
> configuration?
>
> Not sure what you're trying to do, but if you want to run analytics on
> bivio2,
> because you can run single instances of radium on mgt0,  you can have 2
> radium(bivio2).  One radium to collect from bivio1 and offer records to the
> analytics running on the cores, and another to collect records from the
> core
> based analytics to get the processed data (alarms / alerts / summaries)
> to send out of bivio2.
>
> Is that even close to what you would want to do?
>
> Carter
>
> On Jul 19, 2013, at 2:37 AM, CS Lee <geek00l at gmail.com> wrote:
>
> > hi Carter,
> >
> > Alright after boxing with bivio, now I realize why argus(bivio
> 1)->radium(bivio2) -> argus client is not working and why the timeout
> happens.
> >
> > argus(bivio 1) runs fine, but radium(bivio 2) becomes the culprit
> because of how Bivio is designed. Bivio scales the process to multiple
> CPUs, and because it has 12 cores, 12 instances of radium are racing to
> connect to argus on bivio 1, and all these 12 instances of radium are
> getting same copies of data from argus as all of them are hitting 100% CPU
> usage, and because of that as well, argus can't serve anymore client,
> that's why I saw the timeout problem. The other argus client can retrieve
> data when one of radium released the connection again.
> >
> > Radium is not suitable to run on Bivio because of how Bivio hardware is
> designed, and that's very painful to know. Therefore even if I can change
> argus to accept more than 12 instances of client, the issue lies in all the
> radium instances in Bivio are capturing the same data and that makes things
> worse because of duplication and intensive computing resource usage.
> >
> > If anyone has more experience in running radium on Bivio, do let me
> know, I have known some people who runs single instance of radium on mgt0,
> but then what's the point of using Bivio if it is not for scaling purpose.
> >
> > Cheers!
>
>


-- 
Best Regards,

CS Lee<geek00L[at]gmail.com>

http://geek00l.blogspot.com
http://defcraft.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pairlist1.pair.net/pipermail/argus/attachments/20130719/276580b0/attachment.html>