VoIP Mailing List Archives

[michiel at vanbaak.info]

On 15:00, Thu 17 Jan 08, Matt wrote:

[ira at extrasensory.com]

At 11:53 AM 1/18/2008, you wrote:

[burke at tailorhosting...]

[michiel at vanbaak.info]

On 14:34, Fri 18 Jan 08, Ryan Burke wrote:

[ira at extrasensory.com]

At 12:34 PM 1/18/2008, you wrote:

[tzafrir.cohen at xorco...]

On Fri, Jan 18, 2008 at 12:20:56PM -0800, Ira wrote:

[joakimsen at gmail.com]

I'm running 1.4 in production on the following two systems:

Tyan GT20 AMD 939 dual core. openSuSE x86_64 10.1
Celeron 2.4ghz RHEL 4... cheap server from ThePlanet from what I
recall they use cheap cheap cheap consumer grade stuff.

Not a single crash not a single issue.

I will admit we run magnitudes more traffic on 1.2 thusfar but I think
we would have seen issues with 1.4 if they were. FWIW I only started
testing 1.4 approx 6 months ago and put it in production approx 3
months ago.

On Jan 18, 2008 3:20 PM, Ira <ira at extrasensory.com> wrote:

[rob at hillis.dyndns.org]

I would suspect that your hardware is the cause of your problems.
Running a production PBX system on a discarded desktop system is a
/really/ bad idea.

I would seriously consider an upgrade to your hardware.
Ira wrote:

[ira at extrasensory.com]

At 02:20 PM 1/18/2008, you wrote:

[ira at extrasensory.com]

At 03:35 PM 1/18/2008, you wrote:

[rob at hillis.dyndns.org]

What you run it on is very much a function of how reliable you want the
system to be. The better the hardware, the more reliable it will be.
If you're running in a business environment, then I wouldn't recommend
anything less than server grade - even if it's low end server grade.
The company I work for supplies either Dell PowerEdge 860s (1RU servers
that are similar in price to an upper mid-range desktop) or Dell
PowerEdge 840s (tower cases that are similar in price to a mid-range
desktop)

Running on cheap hardware is a great way to cost yourself more in the
long run - in lost productivity, lost sales and IT support.
Ira wrote:

[tzafrir.cohen at xorco...]

On Sat, Jan 19, 2008 at 06:21:15PM +1100, Rob Hillis wrote:

[ira at extrasensory.com]

At 04:44 AM 1/19/2008, you wrote:

Well, it's been very reliable. It's been running 24/7 for 2 years and
the only problems have been my putting bugs in the dial plan,
problems with SIP providers going broke and trying 1.4. So how
exactly would more expensive hardware have improved my reliability?

I really wanted it to be reliable. I ran a duplicate network for the
phones so they don't share bandwidth, I bought good phones.
Everything is POE with the POE switch, the Asterisk box and all the
networking stuff to the outside on a big UPS.

Do you actually think the odds of a HP desktop sitting on a UPS in a
cool corner doing nothing suddenly dying are much greater than a
bottom end server box doing the same? It seems to me unless I want
to go dual PS and flash drives that I'm not going to do much better
than I have now.

Ira

[benny+usenet at amorse...]

Tzafrir Cohen <tzafrir.cohen at xorcom.com> writes:

[rob at hillis.dyndns.org]

I wasn't intending to blame Ira for his own problems - I was intending
to point out that running a production system on discarded hardware is a
really bad idea.

I wasn't even suggesting a mammoth server - as you may or may not have
seen in my subsequent reply to him, the place I work for sells fairly
low-end servers as Asterisk boxes which (at least in Australia) are
comparable to mid to upper-mid range desktops in terms of pricing. 90%
of the serious reliability problems I've seen are on hardware that
people have taken the really cheap route on.

Most people seem to think that Asterisk is a really cheap PBX. While
Asterisk is certainly /cheaper/ than just about all comparable PBXs, if
it's to be done properly and reliably it's certainly not dirt cheap.
Evaluating Asterisk certainly can be since if it's only a test system,
you can scrounge up some older hardware. The real mistake is in putting
the older hardware into full production.
Tzafrir Cohen wrote: