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

20030620: HDS feed to/from seistan (cont.)



>From: Robert Leche <address@hidden>
>Organization: LSU
>Keywords: 200306161954.h5GJs2Ld016710 LDM-6 IDD

Hi Bob,

re:
>>My next test will be to return metlab.cas.usf.edu's HDS feed to
>>pluto.met.fsu.edu where the receipt latencies were near zero.  I
>>would like to then feed the HDS data _to_ seistan to see if the
>>throttling is bidirectional.  In order to really test this, I
>>would need to shut off the HDS feed seistan is currently getting.
>>Is this OK with you?

>Are you saying you want to turn off HDS or do you want to source HDS from
>some  other than our current source?

I wanted to feed HDS data into seistan from a machine here at Unidata.
I then wanted to track the HDS latencies as seen by seistan to see if
the throttling was equal inbound and outbound.  Since the earlier
email, I noticed that sirocco is ingesting HDS from both datoo and
unidata2.ssec.wisc.edu, and that the majority of the HDS data it is
receiving is from unidata2.ssec.wisc.edu.  On top of that, the
latencies for getting the data from unidata2.ssec.wisc.edu were
virtually zero.  This tells me -- assuming that seistan and sirocco are
really on the same subnet and not separated by any routers/switches --
that there is no problem in getting HDS data into SRCC, at least to
sirocco.

>As far as I know, we are not using HDS,
>so removing it from the configuration will not impact LSU. That being the
>case, turn it off.

Thanks.

(a few minutes later)

I have been running 'ldmadmin watch -f HDS' on seistan watching HDS
data come in from emo.unidata.ucar.edu.  It is arriving with little to
no latency proving that there is no problem getting data into seistan.
There still is a problem of getting that same data out of seistan to a
remote host, however.  I am ingesting the HDS feed from seistan on
zero.unidata.ucar.edu, and see that the latencies are running at around
50 minutes:

HDS feed route                                 Latency
----------------------------------------------+------------------------
emo.unidata.ucar.edu -> seistan.srcc.lsu.edu   near zero seconds
seistan.srcc.lsu.edu -> zero.unidata.ucar.edu  50 minutes


The next question I have is whether or not the same effect would be
seen feeding HDS from sirocco to zero.unidata.ucar.edu.  This would
tell us if the problem is isolated to seistan, or if it is associated
with all outbound traffic from either srcc.lsu.edu or lsu.edu.

I can not setup a feed test from sirocco because sirocco is not
allowing feed requests from hosts in the unidata.ucar.edu domain.  I
tried using the same 'root' password on sirocco as I used on seistan,
but it did not work.  Can you give me the root password to sirocco?
(mail it to address@hidden without mention of machine or
account names).  Thanks.

Tom