[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 20001023: New upstream host



Hi Tim, 

We generally to traceroutes and netcheck for network performance as well.

If you are seeing far better connections to Iowa St. than from Arizona
let's see if Geff can allow you to feed from 

homer.geol.iastate.edu

Tim, if this is ok with Geff we will keep nimbus (Arizona) as your
failover and have homer at Iowa State be your primary.

Geff, 

Would it be possibe for you to place an allow UNIDATA for Tim's
machine

hailshaft.atmo.ttu.edu

Thank you and please cc me on future correspondence

-Jeff
____________________________                  _____________________
Jeff Weber                                    address@hidden
Unidata Support                               PH:303-497-8676 
NWS-COMET Case Study Library                  FX:303-497-8690
University Corp for Atmospheric Research      3300 Mitchell Ln
http://www.unidata.ucar.edu/staff/jweber      Boulder,Co 80307-3000
________________________________________      ______________________

On Mon, 23 Oct 2000, Unidata Support wrote:

> >To: address@hidden
> >From: Tim Doggett <address@hidden>
> >Subject: New upstream host
> >Organization: UCAR/Unidata
> >Keywords: 200010231906.e9NJ6c413182
> 
> Hi,
> 
> We have been struggling with severe latency problems for about six months.
> We get timely data in the UNIDATA feed from 10Z to 14Z, at all other hours
> of the day we get latencies greater then an hour (3660 secs +).  The other
> feeds seem to work fine (NLDN, PCWS).  
> 
> My first thought was that we were overloading our ingestor.  We recently
> obtained a SUN Ultra 10 that is dedicated to only data ingesting (no GEMPAK
> gridding, image generation, or other background processes) to replace the
> x86 machine we were using, and the problem persists the same as before.
> When we got the new SUN, I triple checked the installation procedure to
> make sure it was followed to the letter.  I also tried the hack to split
> the feeds by IP address and DNS address, and that does not seem to make a
> noticable difference.  
> 
> I have now turned my attention in the upstream direction.  Right now we are
> feeding off of nimbus.atmo.arizona.edu most of the time, as the A&M feed is
> very hit and miss (I believe we determined it was their firewall).  I have
> spent some time during the last week pinging and traceroute'ing various IP
> addresses listed on the LDM page, and there are three that have
> significantly better connections then Arizona.  They are Iowa State,
> Oklahoma, and Ohio State.  Iowa State was the best, and Ohio State and
> Oklahoma were nearly the same but not a whole lot better then Arizona.
> 
> Is it possible to get Iowa State to allow us (hailshaft.atmo.ttu.edu)?
> Should I contact them directly (if so, who)? If not, do you have any other
> ideas regarding trouble shooting the network?  Thanks a bunch for your time.
> 
> -Tim
>