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

20050318: oliver?

>From: David Wojtowicz <address@hidden>
>Organization: UIUC
>Keywords: 200503181316.j2IDGGv2020338 IDD CONDUIT

Hi David,

>  Why is oliver.unidata.ucar.edu requesting CONDUIT from 
>banff.atmos.uiuc.edu?    The LDM on banff was only established to 
>ingest a small portion of the CONDUIT data from our main LDM server, 
>flood, in order to initialize a model run.

This was my doing.  I wanted to setup up oliver to ingest from the main
machine getting CONDUIT at UIUC.  For some reason, I thought that you
had switched to using banff for your CONDUIT ingest.

>Looks like oliver is 
>taking advantage of the default allow for unidata in the config file.   

Yes, we needed to setup alternate feeds for CONDUIT when
atm.geo.nsf.gov went offline last week.

>   Its causing banff to do 26GB outbound traffic daily for at least the 
>last week.  What's up with this?

We want to have a redundant feed of CONDUIT into idd.unidata.ucar.edu,
and oliver is one of the collector machines for the idd.unidata
cluster.  I just changed the request in oliver to flood.  Sorry for my
mistaking banff's role.


NOTE: All email exchanges with Unidata User Support are recorded in the
Unidata inquiry tracking system and then made publicly available
through the web.  If you do not want to have your interactions made
available in this way, you must let us know in each email you send to us.

>From address@hidden  Fri Mar 18 09:46:43 2005

OK...   Next time it'd be good if you sent a note to inform us of your  
intentions in advance.    We trust your topology decisions, have no  
problem with them and are happy to respond to your requests, but   
CONDUIT is no tiny amount of data to move around and we'd like to at  
least  be aware of it.

When the machine started showing up in the top traffic producer lists  
in our network stats reports with 26GB/day outbound  and we didn't  
think we had done anything to account for it I was afraid that it had  
been compromised and was being used for nefarious purposes.   I even  
saw the open connection back to oliver, but I at first thought it was  
rtstats sending data back.  (as rtstats is an alias that could change)   

  Only later did I realize that oliver was feeding from it.... I could  
have easily discovered this from the logs or by looking up rtstats to  
see that it was not oliver, but it didn't occur to me that there'd be  
any outbound feed from banff and I went looking elsewhere first.

So there won't be future misunderstandings, I'm going to remove the  
carte blanche Unidata allow lines from the various LDMs we have running  
and put them in by your request.  oliver is now allowed to feed conduit  
from flood.

David Wojtowicz, Sr. Research Programmer, Sysadmin
Dept of Atmospheric Sciences / Computer Services
University of Illinois at Urbana-Champaign
address@hidden  (217) 333-8390