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

Re: 20000823: LDM, ldmfail questions from West Chester university



Unidata Support wrote:
> 
> Anne,
> 
> Even though the salutation is aimed at me, the questions being asked are
> really about the LDM and how to setup failing over.
> 
> The comments about the Penn State machine are in reference to my
> comments about it being accessed by UVa as 'navier.meteo.psu.edu'.
> Rajul was told to access it as 'navierldm.meteo.psu.edu'.  Our
> previous exchange is included at the end of the email.
> 
> Tom
> 
> ------- Forwarded Message
> 
> >From: "Pandya, Rajul E." <address@hidden>
> >Subject: RE: 20000823: use of ldm-mcidas for imagery (cont.)
> >Organization: West Chester University
> >Keywords: 200008221228.e7MCSSN28075 ldm-mcidas Unidata-Wisconsin
> 
> Hi Tom,
> 
> Thanks for your help. I thought I was pretty sure that the navierldm link is
> ok - I got it from Jeff Weber and Art Person (at Penn State) just last week.
> It had been fine since then.
> 
> Today, though, I can't get anything from them at all (I attatched a typical
> log file below). On the other hand, when I try the exact same requests to
> snow.cit.cornell.edu (my failover site) everything comes in fine. I did try
> navier.meteo.psu.edu, and had no luck with that either.  The good news is my
> ldmd.conf format was, as you said ok (except for the site). The bad news is
> that something is up with Penn State.

Hi Rajul,  

I can't tell which 'Tom' sent me this - by the time it got to me the address 
was removed.  Would you please forward this to the appropriate 'Tom'?

I would start by evaluating your connection to navierldm - from the log it 
looks at though it's at least up and responding.  We have a web page about this 
- please take a look at: 
http://www.unidata.ucar.edu/packages/ldm/troubleshooting/networkTrouble.html
and give some of those operations a try.  It could be that network trouble is 
causing the connection to go up and down thus not allowing products to get 
through.

> 
> I guess that means I have two questions:
> 
> 1. What should I do about deciding which host to use? I am using the command
> "ldmadmin ldmd.snow.cit.cornell.edu" to get data.

Jeff is away right now, but when he gets back we can discuss whether your 
primary feed should be changed. He should be back tomorrow or Monday.

> 
> 2. How does the ldmfail command work? I suspect I must of set it up
> incorrectly since it didn't catch the trouble at Penn State and
> automatically switch to Cornell. I have disabled the cron job for it for
> now, since it looks like it will automatically try to switch back to Penn
> State.

I suspect failover may not detect a bad network connection - it only pings the 
upstream host.  If the upstream host responds it assumes everything's ok.  ping 
is a very trivial connection, so it may well succeed where a longer duration 
connection would not. 

Let me know how this turns out.


Anne



> I am not sure if this is helpful or not, but here is a typical log file from
> trying to get data from the navierldm site.
> 
> Thanks again for your help.
> Rajul
> 
> Aug 23 20:22:41 rossby.wcupa.edu rpc.ldmd[580]: Starting Up (built: Aug 10
> 1999 15:33:12)
> Aug 23 20:22:41 rossby.wcupa.edu pqexpire[581]: Starting Up
> Aug 23 20:22:41 rossby.wcupa.edu pqbinstats[582]: Starting Up (580)
> Aug 23 20:22:41 rossby.wcupa.edu pqact[583]: Starting Up
> Aug 23 20:22:41 rossby.wcupa.edu navierldm[585]: run_requester: Starting Up:
> navierldm.meteo.psu.edu
> Aug 23 20:22:41 rossby.wcupa.edu navierldm[585]: run_requester:
> 20000823192241.937 TS_ENDT {{WMO,  ".*"},{MCIDAS,  "^pnga2area Q."}}
> Aug 23 20:22:42 rossby.wcupa.edu navierldm[585]:
> FEEDME(navierldm.meteo.psu.edu): reclass: 20000823192241.937 TS_ENDT
> {{IDS|HDS|DDPLUS,  ".*"},
> Aug 23 20:22:42 rossby.wcupa.edu navierldm[585]:
> FEEDME(navierldm.meteo.psu.edu): OK
> Aug 23 20:22:42 rossby.wcupa.edu navierldm[585]: RECLASS: 20000823192242.844
> TS_ENDT {{IDS|HDS|DDPLUS,  ".*"},{MCIDAS,  "^pnga2area Q."}}
> Aug 23 20:22:43 rossby.wcupa.edu navierldm[585]: Connection reset by peer
> Aug 23 20:22:43 rossby.wcupa.edu navierldm[585]: Disconnect
> Aug 23 20:22:43 rossby.wcupa.edu localhost[593]: Connection from localhost
> Aug 23 20:22:43 rossby.wcupa.edu localhost[593]: Connection reset by peer
> Aug 23 20:22:43 rossby.wcupa.edu localhost[593]: Exiting
> Aug 23 20:23:13 rossby.wcupa.edu navierldm[585]: run_requester:
> 20000823192313.056 TS_ENDT {{WMO,  ".*"},{MCIDAS,  "^pnga2area Q."}}
> Aug 23 20:23:13 rossby.wcupa.edu navierldm[585]:
> FEEDME(navierldm.meteo.psu.edu): reclass: 20000823192313.056 TS_ENDT
> {{IDS|HDS|DDPLUS,  ".*"},
> Aug 23 20:23:13 rossby.wcupa.edu navierldm[585]:
> FEEDME(navierldm.meteo.psu.edu): OK
> Aug 23 20:23:13 rossby.wcupa.edu navierldm[585]: RECLASS: 20000823192313.864
> TS_ENDT {{IDS|HDS|DDPLUS,  ".*"},{MCIDAS,  "^pnga2area Q."}}
> Aug 23 20:23:17 rossby.wcupa.edu navierldm[585]: Connection reset by peer
> Aug 23 20:23:17 rossby.wcupa.edu navierldm[585]: Disconnect
> and it goes on like that for a while.
> 
> > -----Original Message-----
> > From: Unidata Support [SMTP:address@hidden]
> > Sent: Wednesday, 23 August 2000 11:07
> > To:   Pandya, Rajul E.
> > Cc:   address@hidden
> > Subject:      20000823: use of ldm-mcidas for imagery (cont.)
> >
> > >From: "Pandya, Rajul E." <address@hidden>
> > >Organization: UCARWest Chester University
> > >Keywords: 200008221228.e7MCSSN28075 ldm-mcidas Unidata-Wisconsin
> >
> > Rajul,
> >
> > >Thanks for the suggestion.  I tried changing my ldmd.conf file from
> > >request UNIDATA ".*" navierldm.meteo.psu.edu
> > >to
> > >request WMO ".*" navierldm.meteo.psu.edu
> > >request MCIDAS "^pnga2area Q." navierldm.meteo.psu.edu
> >
> > Hmm...  Another site feeding from Penn State is using:
> >
> > request MCIDAS "^pnga2area Q[01]" navier.meteo.psu.edu
> >
> > Did Penn State change the name of their machine?
> >
> > A 'nslookup' on navier.meteo.psu.edu gives the expected results:
> >
> > Non-authoritative answer:
> > Name:    navier.meteo.psu.edu
> > Address:  128.118.28.2
> >
> > but the same on navierldm.meteo.psu.edu hangs.
> >
> > >and did ldmadmin stop then start. I let it run for about half an hour and
> > >got WMO data, but no satellite data. I switched it back and got satellite
> > >data right away. Did I do something wrong?
> >
> > Your request line looks fine except for the specification of
> > navierldm.meteo...
> >
> > >My pqact file is all pnga2area commands.
> >
> > OK.  The entries in pqact.conf are just for processing products out of
> > the LDM's queue.  They really don't have anything to do with the ldmd.conf
> > request entries.
> >
> > What happens when you change to use of navier.meteo.psu.edu in your
> > request
> > line?
> >
> > Tom
> > **************************************************************************
> > Unidata User Support                                    UCAR Unidata
> > (303)497-8644                                                  P.O. Box
> > address@hidden                                   Boulder, CO
> > --------------------------------------------------------------------------
> > Unidata WWW Service                        http://www.unidata.ucar.edu/
> > **************************************************************************
> 
> ------- End of Forwarded Message

-- 
***************************************************
Anne Wilson                     UCAR Unidata Program            
address@hidden                 P.O. Box 3000
                                  Boulder, CO  80307
----------------------------------------------------
Unidata WWW server       http://www.unidata.ucar.edu/
****************************************************