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

Re: 20001023: Failover site request from PSU



Hi Jeff, 

Sounds like we have exhausted solutions on the networking end.

I will gather some possible candidates (I2 sites) and have you execute
some traceroutes to determine best connect.

Back soon,

-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 Tue, 24 Oct 2000, Jeff Wolfe wrote:

> In message <address@hidden>
> , Jeff Weber writes:
> > Hi Jeff,
> > 
> > It is against general policy to feed the same institution from two sites.
> > The reason for this is.. you guessed it.. bandwidth. If you also start
> > pulling the data from another machine into your internal network, it may
> > actually make things worse. We STRONGLY discourage this procedure. Other
> > options may be relocating your primary machine to be closer to navier
> > (electronically) to eliminate the feed problems. If the "other I2" site
> > was solely for failover things would probably not get to bottlenecked at
> > PSU if diligence was used to maintain feeding from navier anytime it was
> > up and recieving data. If you are experiencing steady problems from navier
> > and stayed on failover for any duration I believe it could cause severe
> > IDD congestion converging at PSU making your problem worse. Sooo..
>  
> 
> 
> > First lets see if you can contact your sys admin and see if there is a way
> > to improve the hops from your machine to navier, then if nothing can be
> > improved there we may want to find another I2 site for failover. It is
> > "policy" to have only 1 feed to an institution and then having that
> > machine feed ALL internal machines, but PSU has been great at providing
> > access to other institutions so we want to make sure you all are happy and
> > have the data you need in a timely manner. I would prefer to solve this
> > from an internal network position, but remain open to make sure your needs
> > are met..
> > 
> > Let me know,
> 
> 
> Jeff, 
> 
> The Issue is more one of failover. We have some people ingesting SRUS 
> bulletins who really need continuous uninterrupted access to the data stream. 
> The reliability issues I'm worried about have to do with reliability of the 
> server "navier.meteo.psu.edu" itself.. Under normal operations, I don't 
> anticipate ever feeding from anywhere other than navier. I'm interested in 
> having perhaps one backup site for situations like last night when navier's 
> ldm was unavailable for several hours. Topologically, navier and catena are 
> about as close as they can be on our network and local network connectivity 
> is 
> not an issue.
> 
> -Jeff
> 
>