Re: [Wg-test-framework] [Minutes] March Test WG meeting

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Slutz, Donald Christopher writes ("Re: [Wg-test-framework] [Minutes] March Test 
WG meeting"):
> I just got off the phone with EarthLink.  A rack can have 2x 30A/208 power.
> Looks like 1 30A/208 would fit.  We are limited to 24A steady draw.  So
> 24*208 is less than 27*200, but if 44 for many of the servers is right;
> we should be fine (more cost for 2x 30A).

I'm not sure what "30A/208" means.  Is that 208V ?  I hadn't
appreciated that US datacentres use electricity which is nearly as
good as European electricity :-).

We should budget based on 200VA per server, because those are
measurements of non-idle XenRT test boxes, and because that's what I
have based my Jump costings on.

> On the network bandwidth cost, they have changed the way they handle this.
> They now use "95th percentile bandwidth metering". I found
> http://www.semaphore.com/blog/2011/04/04/95th-percentile-bandwidth-metering-explained-and-analyzed
> to be an easy read.  Still have no idea what our network bandwidth should
> be.  For now I have it at 2Mbps.

I think we should budget 4Mbps.  We don't have any useful measurements
of the total bandwidth use at this stage and they would be difficult
to get.

My own colo uses about 1Mbps and that includes serving the test logs
(which are currently being hosted there) but that figure does not
include inbound data (eg downloads of source tarballs etc.)

Don, will you have a complete costing, along the lines of the one I
just sent for Jump, available by the meeting ?

Ian.

_______________________________________________
Wg-test-framework mailing list
Wg-test-framework@xxxxxxxxxxxxxxxxxxxx
http://lists.xenproject.org/cgi-bin/mailman/listinfo/wg-test-framework