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

Re: 20010426: New LDM/IDD node

Hi Adam, 

That is quite a list of requests.

Are you using all the data you are requesting?

You will see optimal performance when the load is lighter.

Given the fact you are seeing 30 min latencies over such a short connect,
we may be hammerring your machine or the network. With the ..If i
remember...dual p3 933's and a Gig of RAM I suspect it is not your
machine.

Maybe LSU was down and we are catching up, has the latency been steadily
at 30 mins or have you been aware of it in the past. If LSU were to go
down for a period, products would be late for awhile until the LDM caught
up..If the latency persists at 30 mins we would want to evaluate your
requests and narrow them down if possible. 

For example, are you using the NOGAPS model output...by the way that
should be point to point from FNMOC like the NLDN data from albany.

LCDR David

Dimitriou

address@hidden

831.656.4350

from the URL:

http://unidata.ucar.edu/projects/idd/sitelist.html

The GEM data is not available yet, nor is NPORT, EXP is being used
sporadically for field experiments, GPS is used only for participating
SuomiNet sites.

Soooo, without really knowing what data you are using, I would suggest
changing the line to UNIDATA|NNEXRAD|FSL2 and the FSL2 only if you are
planning on using the profiler data.

Looks like you have limited your NNEXRAD feed to 4 sites so that is good.

We have also found that on some OS's it is not OK to have multiple request
lines to the same host, it reads the first, and reclasses other s after
that.

Can you send me the last few lines of your log file..

I am beginning to get a better picture of what may be going on.

Can be a bit painful for awhile, but we'll get it hummin'

Thank you,


-Jeff
____________________________                  _____________________
Jeff Weber                                    address@hidden
Unidata Support                               PH:303-497-8676 
NWS-COMET Case Study Library                  FX:303-497-8690
University Corp for Atmospheric Research      3300 Mitchell Ln
http://www.unidata.ucar.edu/staff/jweber      Boulder,Co 80307-3000
________________________________________      ______________________

On Thu, 23 Aug 2001, adam  taylor (994) wrote:

> The latency is about 30 minutes between when the time on the file and the
> time we get it.  That i know is not good.  We are requesting:
> 
> request UNIDATA|GEM|FAA604|NPORT|EXP|GPS|FSL|NOGAPS   ".*"    sir(etc...)
> request NNEXRAD "/p...(SHV|JAN|LZK|POE)"                      sir(etc...)
> request NLDN    ".*"    striker.atmos.albany.edu
> 
> I just realised that we are not getting the NNEXRAD from sirocco but from
> thelma.ucar.edu.  That would explane why I only got NNEXRAD for long
> periods of time today.  Would that be the reason for the long latency??
> The fact that two servers are connected at the same time.  We are also
> connected to striker.atmos.albany.edu for the NLDN only. thelma is for the
> NNEXRAD only.  I'm switching the thelma one back to LSU now.  The correct
> ones now are listed above.
> 
> Thanks again
> 
> Adam
> ___________________________________________________________________________
> 
> 
> On Thu, 23 Aug 2001, Jeff Weber wrote:
> 
> > HI Adam, 
> > 
> > One more thing, when you do get your data, what is the latency of the
> > products? If near none, may not be a network issue, if pushing an hour, we
> > may have located the problem...
> > 
> > Thanks,
> > 
> > -Jeff
> > ____________________________                  _____________________
> > Jeff Weber                                    address@hidden
> > Unidata Support                               PH:303-497-8676 
> > NWS-COMET Case Study Library                  FX:303-497-8690
> > University Corp for Atmospheric Research      3300 Mitchell Ln
> > http://www.unidata.ucar.edu/staff/jweber      Boulder,Co 80307-3000
> > ________________________________________      ______________________
> > 
> > On Thu, 23 Aug 2001, Jeff Weber wrote:
> > 
> > > Hi Adam,
> > > 
> > > (jweber) laraine:/home/jweber 1 % ldmping -i5 -h sirocco.srcc.lsu.edu
> > > Aug 23 19:55:33      State    Elapsed Port   Remote_Host
> > > rpc_stat
> > > Aug 23 19:55:33 RESPONDING   0.127509  388   sirocco.srcc.lsu.edu  
> > > Aug 23 19:55:38 RESPONDING   0.036370  388   sirocco.srcc.lsu.edu  
> > > Aug 23 19:55:43 RESPONDING   0.035674  388   sirocco.srcc.lsu.edu  
> > > Aug 23 19:55:48 RESPONDING   0.035661  388   sirocco.srcc.lsu.edu  
> > > Aug 23 19:55:53 RESPONDING   0.035857  388   sirocco.srcc.lsu.edu  
> > > Aug 23 19:55:58 RESPONDING   0.035385  388   sirocco.srcc.lsu.edu  
> > > Aug 23 19:56:03 RESPONDING   0.035772  388   sirocco.srcc.lsu.edu  
> > > Aug 23 19:56:08 RESPONDING   0.035702  388   sirocco.srcc.lsu.edu  
> > > 
> > > 
> > > Hmmm, cannot replicate here. This lends me to believe it may have
> > > something to do with your ISP or network conditions. Perhaps it is getting
> > > "clogged" and therefore cannot resolve the ldmping packets.
> > > 
> > > The NNEXRAD issue may also be playing a part, we (LSU) may be feeding too
> > > many folks the NNEXRAD feed, so the machine may fall behind for awhile
> > > during these peak data periods...When this happens again, try doing a
> > > traceroute, I know these are small packets, but it may lend some
> > > additional info.. 
> > > 
> > > 
> > > There is no reason the LDM should just quit, and I am very surprised there
> > > is no log entry regarding the shutdown. Could you please snip the tail of
> > > your log and send to me, maybe I can see something in there that might
> > > give a clue as to what is going on.
> > > 
> > > I will contact LSU and see if any other downstream sites have mentioned
> > > difficulties, we have not heard from them here at Unidata..but that does
> > > not always mean all is well.
> > > 
> > > What are you requesting from LSU?
> > > 
> > > How is your request line configured?
> > > 
> > > I guess we still have some "tuning" left for your LDM..
> > > 
> > > 
> > > I am cc'ing Anne Wilson who helps develop the LDM code, she is away this
> > > week, but will be back monday and can lend insight then if we have not
> > > solved the problem by then.
> > > 
> > > Thank you,
> > > 
> > >  -Jeff
> > > ____________________________                  _____________________
> > > Jeff Weber                                    address@hidden
> > > Unidata Support                               PH:303-497-8676 
> > > NWS-COMET Case Study Library                  FX:303-497-8690
> > > University Corp for Atmospheric Research      3300 Mitchell Ln
> > > http://www.unidata.ucar.edu/staff/jweber      Boulder,Co 80307-3000
> > > ________________________________________      ______________________
> > > 
> > > On Thu, 23 Aug 2001, adam  taylor (994) wrote:
> > > 
> > > > I am the new contact.  address@hidden
> > > > 
> > > > I have an interesting problem on this end but I'm not sure who to direct
> > > > this to.  When i run:
> > > > 
> > > > ldmping -i5 -h sirocco.srcc.lsu.edu
> > > > 
> > > > I get:
> > > > 
> > > > (DATE)     State Elapsed  Port Remote_Host          rpc_stat
> > > > (THE DATE) NAMED 12.132903   0 sirocco.srcc.lsu.edu can't contact
> > > >                                                     portmapper: 
> > > >                                                     RPC: Timed out 
> > > > and i get that almost without fail every time.  Every once and a 
> > > > while i get it to say RESPONDING and a time of around 3 or so.
> > > > 
> > > > I have gotten numbers in the 30's before but for the most part, 12 or so
> > > > is the average.
> > > > 
> > > > Also, I have been watching LSU's feed for a while and there are 5 to ten
> > > > minute times where i get nothing but nexrad feed products and even then 
> > > > it
> > > > is real slow.
> > > > 
> > > > Also one more thing, the ldm server is just spontaniously quiting with 
> > > > no
> > > > error message of anykind.  I have noticed that it happens, or seems, 
> > > > then
> > > > when there is a really bad network lag it seems to quit.  I was watching
> > > > it today and just all of a sudden the feed stopped.  when i ran ldmadmin
> > > > isrunning it came back as no.  I have version 5.1.2 running on a dual P3
> > > > 933 with 1 GIG RAM running RedHat 7.0
> > > > 
> > > > Any throughts on this???
> > > > 
> > > > Thanks in advance
> > > > 
> > > > Adam
> > > > _________________________________________________________________________
> > > > 
> > > > On Thu, 23 Aug 2001, Jeff Weber wrote:
> > > > 
> > > > > Hey Adam, 
> > > > > 
> > > > > While I have you...
> > > > > 
> > > > > Is 
> > > > > 
> > > > > Chris Clarke
> > > > > 
> > > > > address@hidden
> > > > > 
> > > > > 318-342-1891
> > > > > 
> > > > > Still current contact info for UNL or should you  be the contact?
> > > > > 
> > > > > Thank you,
> > > > > 
> > > > > -Jeff
> > > > > ____________________________                  _____________________
> > > > > Jeff Weber                                    address@hidden
> > > > > Unidata Support                               PH:303-497-8676 
> > > > > NWS-COMET Case Study Library                  FX:303-497-8690
> > > > > University Corp for Atmospheric Research      3300 Mitchell Ln
> > > > > http://www.unidata.ucar.edu/staff/jweber      Boulder,Co 80307-3000
> > > > > ________________________________________      ______________________
> > > > > 
> > > > > On Thu, 23 Aug 2001, adam  taylor (994) wrote:
> > > > > 
> > > > > > Ok, the feed has decided to return to normal, we are finally 
> > > > > > getting the
> > > > > > HDS and WMO feeds too.  Thanks anyways.
> > > > > > 
> > > > > > Adam
> > > > > > 
> > > > > > 
> > > > > 
> > > > > 
> > > > 
> > > > 
> > > 
> > > 
> > 
> > 
> 
>