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

[LDM #WXT-754815]: Blocked by Jeff Weber

Hi Grant,

re: was there anything unusual going on at DRI recently?

> Over the weekend, Saturday July 21, DRI was having some networking
> connectivity issues. I do not know exactly what the issue was, but we
> received multiple reports that our websites were unreachable.

OK.  We saw the problem with LDM connections from DRI yesterday (Wednesday),
not over the weekend.

re:
> We are
> running a bit of an older LDM version, 6.5.1, are there known issues when
> there are connectivity problems?

No.  That being said, it would be a good idea to upgrade to a recent version
of the LDM.  Steve Emmerson (the LDM developer here) is working on a new
version, so you will want to wait for that one (v6.10.2) before upgrading
if you want to upgrade to v6.10.x.

re:
> Unfortunately our LDM logs only go back about 24 hours, otherwise I would
> make them available to you.

OK, thanks for looking.

re:
> We do not request CONDUIT ".*", but here is what our CONDUIT requests
> looked like until yesterday:
> 
> request        CONDUIT         "(nam.*awip3d*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(nam.*awip12*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(nam.*awphys*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf0*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf1*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf2*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf3*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf4*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf5*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf6*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf7*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf8*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrbf9*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f0*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f1*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f2*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f3*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f4*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f5*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f6*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f7*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f8*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(gfs.*pgrb2f9*)" thelma.ucar.edu         
> PRIMARY
> request        CONDUIT         "(MT.gfs_CY|MT.nam.*#212)" thelma.ucar.edu     
>     PRIMARY

Is the bandwidth into your shop so small that you determined that you needed
to split your feed this finely?

re: 
> And as of yesterday, to reduce our number of connections, we have changed
> to:
> 
> request CONDUIT         "(nam.*awip3d*)"           idd.unidata.ucar.edu
> request CONDUIT         "(nam.*awip12*)"           idd.unidata.ucar.edu
> request CONDUIT         "(nam.*awphys*)"           idd.unidata.ucar.edu
> request CONDUIT         "(gfs.*pgrbf[012]*)"       idd.unidata.ucar.edu
> request CONDUIT         "(gfs.*pgrbf[345]*)"       idd.unidata.ucar.edu
> request CONDUIT         "(gfs.*pgrbf[6789]*)"      idd.unidata.ucar.edu
> request CONDUIT         "(gfs.*pgrb2f[01234]*)"    idd.unidata.ucar.edu
> request CONDUIT         "(gfs.*pgrb2f[56789]*)"    idd.unidata.ucar.edu
> request CONDUIT         "(MT.gfs_CY|MT.nam.*#212)" idd.unidata.ucar.edu

This is much better from our point of view.  Each connection REQUEST
results in a separate rpc.ldmd (LDM versions < 6.10.x)/ldmd (LDM
versions 6.10.x+) invocation.  Each one consumes resources on our
end, so it is best when downstreams (like you) can limit the number
of REQUESTs.

Question:

- can the pgrbf and pgrb2f requests be further consolidated without
  affecting product receipt latency?

re:
> I hope this is helpful. Let me know if there's anything more I can do.

Thanks, this is helpful.

Cheers,

Tom
--
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
(303) 497-8642                                                 P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage                       http://www.unidata.ucar.edu
****************************************************************************


Ticket Details
===================
Ticket ID: WXT-754815
Department: Support IDD
Priority: Normal
Status: Open