Re: {bb} bb redundancy

[Philip Clark <bblist@cannae.uklinux.net>]

First, the disclaimer :) Your installation is much larger than
anything I've worked with, so anything I can say is based on a
combination of extrapolation, theory and guess-work. Hopefully,
others with similarly sized set-ups will be able to provide more
detailed advice.

I expect that the load average on your DISP/NET machines is
not a problem, unless they are doing additional work. There
should be plenty of processor power there. Similarly, 1GB of
memory might seem low for those processors, but I wouldn't
expect you to be seeing much, if any, swap activity. I'm
going to assume that all three machines have at least 100Mb
network cards. If your DISP/NET machines don't have gigabit
networking, the bandwidth is probably limiting them.

Now, on your options for redundancy. The simplest route would
be to identify a pair of machines that are doing other jobs
and can act as failover BBNET machines. Under normal circumstances,
the extra load on these will be negligible, but they should be on
a similar scale to your current BBPAGER and not be heavily loaded.

If you can dedicate another machine (similar to your current
BBPAGER) to the task, you could have a different BBPAGER
machine for each side of your installation and run them as
failover BBNET machines. It would be possible for your current
BBPAGER to act as failover BBNET for both sides, but I think
that this would require either significant script mangling,
or migrating one half of your installation to a non-standard
port. Neither of those routes appear particularly attractive.

Cheers, Phil.



On Wed, 2005-06-29 at 13:51, Jason Beltrame wrote:
> There are roughly 400 servers per bbdisplay, with a total of 2500
> checks per bbdisplay ( includes client checks )  They both run on dual
> 2.4 xeon's with 1gb of mem.  The current bbpager is just a dual 450. 
> All run freebsd.  Sorry for not including that earlier.
> 
> Thanks,
> 
> J



-- 
But in our enthusiasm, we could not resist a radical overhaul of the
system,
in which all of its major weaknesses have been exposed, analyzed, and
replaced
with new weaknesses. (Bruce Leverett, "Register Allocation in Optimizing
Compilers")

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-==-=-=-=-=-=-=-=-=-=-=-=
To unsubscribe from this list, or to subscribe to the bb-digest list
send e-mail to mailto:majordomo@bb4.com with unsubscribe bb -and/or-
subscribe bb-digest in the BODY of the message.