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

[IDD #BBT-775736]: Server replacement

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: [IDD #BBT-775736]: Server replacement
  • Date: Wed, 05 Sep 2012 18:23:45 -0600
Hi David,

re: testing of a new LDM machine

> We have had the new system configured to read data from the current system 
> during the
> development phase.  We are now entering the test and evaluation phase and 
> were hoping to
> have them work independently since the current system will be retired as the 
> new one
> becomes operational.  This would help us verify that the new system is 
> working with its
> final configuration.

I am not sure what additional information can be gleaned by two machines 
feeding from
the same upstream site simultaneously.  As I understand things, both machines 
are on
the same subnet; presumably their REQUESTs will go through the same firewall 
setup
at your site; and the processing that will be done is independent on the source
of the data feeds.  If all of this is true, the only difference is the source of
the data.

re:
> If that is not possible, then we will have to modify our test and
> evaluation plan and do as you suggest below.

The problem on our end is that the ~ldm/etc/ldmd.conf ALLOW needs to be changed
on all of the real server backends of the idd.unidata.ucar.edu cluster, and
the LDMs running on those systems need to be restarted.  The LDM restart
on each cluster node causes 100-150 downstream feeds to be closed.  Depending
on how large the LDM queue is (they now vary from 12-20 GB), the restart process
can take enough time that the feed REQUESTs may get redirected to other cluster
nodes.  If the sequencing of restarts of the LDMs is not done with care,
an imbalance in system loads can occur.  Because of this we prefer to not
make small ALLOW changes very often.

If you can convince us that there is a significant benefit to your
having a parallel feed setup in your testing/evaluation plan we will
make the change.  Otherwise, we will implement the change just before
your transition from one machine to the other.

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: BBT-775736
Department: Support IDD
Priority: Normal
Status: Closed