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

20040131: LDM test feed from emo...

>From: Mike Voss <address@hidden>
>Organization: SJSU
>Keywords: 200402010010.i110ACp1023488 IDD test

Hi Mike,

>I started feeding storm.sjsu.edu from emo Saturday(today) about 00Z to
>test our bandwidth, I have a few related questions.

It is a good thing you didn't try before this, because the power to
the entire first floor of our building was off for most of the day.

>- Comment, so far I'm able to get all the data including CONDUIT and
>CRAFT, so I'm psyched about that. This is very convincing evidence to our
>network engineers who now believe me that we have a problem on campus. 

Excellent.

>The current test is running from just inside the campus firewall, I can't
>get nearly the bandwidth from the College of Science.B
>
>- Is there a way to get stats, should the stats be showing up on the
>rtstats page, or will this not work with emo?

You sould see stats from storm.sjsu.edu IF it is reporting them back to
us.  The reqporting requires that you have a rtstats line in storm's
ldmd.conf file and that port 388 traffic back to us is not being
blocked.  In addition, if a 'uname -n' on storm does not return a
fully qualified hostname, then you would have had to edit
~ldm/bin/ldmadmin and set the hostname in it:

change:

chop($hostname = `uname -n`);
# $hostname = "your.hostname.here";

to:

# chop($hostname = `uname -n`);
$hostname = "storm.sjsu.edu";

If you find that you have to make this change, you will need to
stop and restart your LDM for the change to take effect:

ldmadmin stop
ldmadmin start

>- My feed went down from 04Z until 14Z today, was emo fine during this
>period?

I brought the LDM up on emo yesterday afternoon at 15:15 our time.  A
quick check of the stats for it show that it has been up ever since:

http://my.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?IDS|DDPLUS+emo.unidata.ucar.edu

>- I would also like to test the ability to feed someone off campus, is
>there anyway I could feed everything to a machine at Unidata for awhile,
>or should I broadcast this request to the community?

I setup a machine here in the UPC to feed from you.  As a stress test,
I am feeding everything with requests split 11 ways:

request IDS|DDPLUS|FSL2|UNIWISC ".*" storm.sjsu.edu
request HRS ".*" storm.sjsu.edu
request NIMAGE ".*" storm.sjsu.edu
request FNEXRAD|NNEXRAD ".*" storm.sjsu.edu
request CRAFT ".*" storm.sjsu.edu
request NLDN ".*" storm.sjsu.edu
request CONDUIT   "[05]$" storm.sjsu.edu
request CONDUIT   "[16]$" storm.sjsu.edu
request CONDUIT   "[27]$" storm.sjsu.edu
request CONDUIT   "[38]$" storm.sjsu.edu
request CONDUIT   "[49]$" storm.sjsu.edu

Data is flowing, and the machine, dhcp22.unidata.ucar.edu is reporting
stats that are viewable through our realtime stats web pages:

http://my.unidata.ucar.edu/content/software/idd/rtstats/siteindex.php?dhcp22.unidata.ucar.edu

So, the data path is:

UPC NOAAPORT ingesters -> emo -> storm -> dhcp22

We should be able to get a good handle for the feed back to dhcp22
later today.

>Thanks for your help as usual,

No worries.

Tom
--
NOTE: All email exchanges with Unidata User Support are recorded in the
Unidata inquiry tracking system and then made publically 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 Sun Feb  1 11:13:45 2004

Tom,

I think your power outage did explain my outage yesterday. I started the
feed 00Z Saturday or 5PM Friday your time, and I was out from 12Z to 22Z,
which is 5am to 3pm your time Saturday. Anyhow, I'm glad that's all it was
and I apologize for starting my feed without informing you.

The stats should be flowing now, I just had a typo in my rtstats line in
my ldmd.conf.

To be consistent, I used your exact same feed patterns and just changed
storm.sjsu.edu to emo....:

request IDS|DDPLUS|FSL2|UNIWISC ".*" emo.unidata.ucar.edu
request HRS ".*" emo.unidata.ucar.edu
request NIMAGE ".*" emo.unidata.ucar.edu
request FNEXRAD|NNEXRAD ".*" emo.unidata.ucar.edu
request CRAFT ".*" emo.unidata.ucar.edu
request NLDN ".*" emo.unidata.ucar.edu
request CONDUIT   "[05]$" emo.unidata.ucar.edu
request CONDUIT   "[16]$" emo.unidata.ucar.edu
request CONDUIT   "[27]$" emo.unidata.ucar.edu
request CONDUIT   "[38]$" emo.unidata.ucar.edu
request CONDUIT   "[49]$" emo.unidata.ucar.edu

So, this is great, thanks for pulling a feed from storm, we'll see how it
does.

As a stress test, I do plan on stopping and starting the LDM a few times
to see how we'll it can catch up. This machine is not very special (1.4
Ghz, 1 GB RAM), but hopefully it will do for a test)

Thanks!!

-Mike