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

Re: 20030203:changing idd relay machine

comments embedded within text..

-Jeff
____________________________                  _____________________
Jeff Weber                                    address@hidden
Unidata Support                               PH:303-497-8676 
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, 3 Feb 2003, David Knight wrote:

> 
> Hi Jeff,
> > 
> > Hi David, 
> > 
> > Congrats on the new machine!
> Thanks. 
> > 
> > What we often suggest is bringing up the new machine "gusher", and have it
> > run in parallel for a time with redwood. Once you are confident that
> > gusher is stable and getting all feeds desired, notify all those that feed
> > from you (from your ldmd.conf) to change to gusher from redwood. This will
> > allow some flexibility for the transition as both redwood and gusher can
> > offer the feed. Then notify the downstream folks of the "drop dead"
> > date/time for redwood and all should be feeding from gusher by that time.
> 
> We have gusher running, right now it is getting its feed from redwood.
> 
> What I'm curious about is how to handle the upstream sites. 
> Should I ask them to feed both redwood and gusher?

Yes, this is best..(in my opinion unless it is going to be months before
the changeover) That way you can be sure of your upstream feed going to
gusher and not being fed from redwood without looking over your logs to
examine who is getting what from where...I am sure both would work fine,
but would recommend having snow feed both redwood and gusher for a brief
period of time..(1 week or so is generally enough time to get people off
one machine and on another, it is amazing how fast people can act when
their feed goes away ;)


> Or, can I have gusher and redwood feed off from each other,
> then when the upstream site changes their ldmd.conf data will
> still flow to both redwood and gusher?

This is also a possibility, but I prefer the former, but you may have
internal needs that I am unaware of..

> The later seems kinder to our upstream sites, but I'm not sure
> how having machines cross feed from each other will work.

Alot of CPU overhead, may not be a problem..but it might.

> I assume that no data will actually get passed back and forth,
> since the products will already exist in the queue, but there
> still could be significant overhead. I guess I could just
> give it a try...
> > 
> > Thats how we do it.
> > 
> > I am assuming this does not affect "striker"...;)
> 
> No, but striker is also getting long in the tooth. We have
> been waiting for a Unidata/NSF equipment grant program.
> But I'm beginning to wonder if there will ever be another one of these.
> Seems like it has been a few years!
> 

Had a meeting about these this last Friday, it appears as if we can
ressurect these grants..I will keep you posted!


> > 
> > Cheers,
> > 
> > -Jeff
> > ____________________________                  _____________________
> > Jeff Weber                                    address@hidden
> > Unidata Support                               PH:303-497-8676 
> > 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, 3 Feb 2003, Unidata Support wrote:
> > 
> > > 
> > > Reply-to:David Knight <address@hidden>
> > > --------
> > > 
> > > ------- Forwarded Message
> > > 
> > > >To: address@hidden
> > > >cc: address@hidden
> > > >From: David Knight <address@hidden>
> > > >Subject: chaning idd relay machine
> > > >Organization: UCAR/Unidata
> > > >Keywords: 200302031501.h13F18627520
> > > 
> > > Hi,
> > >     The time has come to change the machine we use for our IDD relay.
> > > To insure continuity of both our local data, and our downstream sites,
> > > we will probably want to run both idd machines in parallel for some time.
> > > 
> > > Is there a "best practices" writeup somewhere on the recommended procedure
> > > for transitioning from one IDD machine to another? I searched but
> > > could not find anything - maybe I didn't search hard enough...
> > > 
> > > Here is what I was thinking. Lets say we have the following setup
> > > snow - our primary upstream IDD data source
> > > redwood - our old IDD relay
> > > gusher - our new IDD relay
> > > 
> > > Can I set up redwood so that it makes requests to both snow and gusher?
> > > Similarly, can I set up gusher to make requests to both snow and redwood?
> > > Then ask snow to change their allow statements and be assured of
> > > continuous data feeds into both redwood and gusher? It seems like this
> > > should work, but, I want to make sure we are not missing anything.
> > > 
> > > Thanks
> > > David
> > > 
> > > David Knight
> > > Department of Earth and Atmospheric Sciences   Tel: (518)-442-4204
> > > University at Albany   ES-228                  Fax: (518)-442-4494
> > > Albany, NY  12222                              Email: address@hidden
> > > 
> > > 
> > > ------- End of Forwarded Message
> > > 
> > > 
> > 
> > 
>