[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
20030204: McIDAS - Tru64 - ptlist not listing CMAN stations (cont.)
- Subject: 20030204: McIDAS - Tru64 - ptlist not listing CMAN stations (cont.)
- Date: Wed, 05 Feb 2003 09:15:59 -0700
>From: Angel Li <address@hidden>
>Organization: University of Miami
>Keywords: 200301211600.h0LG0gD13023 McIDAS-XCD CMAN
Hi Angel,
>I'm back to this project now... We use a small script to dump a CMAN
>station's data:
>
> #!/bin/sh
>
> MCHOME=$HOME
> MCDATA=$MCHOME/workdata
> MCPATH=${MCDATA}:$MCHOME/data:$MCHOME/help
> MCGUI=$MCHOME/bin
> MCTABLE_READ="${MCDATA}/MCTABLE.TXT;${MCHOME}/data/ADDESITE.TXT"
> MCTABLE_WRITE="/mcidas/data/ADDESITE.TXT"
> MCCOMPRESS=TRUE
> PATH=${PATH}:$MCGUI
> export MCHOME MCDATA MCPATH MCGUI MCTABLE_READ MCTABLE_WRITE MCCOMPRESS
> PATH
>
> STATION=$1
> FILE=$2
>
> mcenv <<EOF
> ptlist.k RTPTSRC/SHIPBUOY.$FILE PAR=DAY TIME "T[F]" "STI[F]"
>"PSL[MB]" DIR "SPD[KTS]" SEL="TYPE CMAN; IDA $STATION" NUM=ALL
> exit 0
> EOF
>
>and if I invoke this script as
>
> % ./cman_dumper0.sh MLRF1 0
>
>the ptlist does not know about the any of the new stations defined even
>though stnlist.k knows about them.
I logged onto sapodilla as 'mcidas' to poke around and see if I could
figure anything out. What I found was what I will call a classic
gotcha with McIDAS ADDE:
The DATALOC for the dataset RTPTSRC was set to point at the ADDE server
at the University of Georgia, not your own machine:
mcidas* dataloc.k LIST
Group Name Server IP Address
-------------------- ----------------------------------------
AMRC UWAMRC.SSEC.WISC.EDU
CIMSS <LOCAL-DATA>
GILMORE GILMORE.SSEC.WISC.EDU
GINICOMP CACIMBO.GGY.UGA.EDU
GINIEAST CACIMBO.GGY.UGA.EDU
GINIWEST CACIMBO.GGY.UGA.EDU
LOCALGIL <LOCAL-DATA>
ME7 IO.SCA.UQAM.CA
MYDATA <LOCAL-DATA>
NEXRCOMP <LOCAL-DATA>
RTGRIDS PAPAGAYO.UNL.EDU
RTIMAGES <LOCAL-DATA>
RTNEXRAD <LOCAL-DATA>
RTPTSRC CACIMBO.GGY.UGA.EDU
RTWXTEXT WEATHER2.ADMIN.NIU.EDU
TOPO <LOCAL-DATA>
WALLOPS WALLOPS.SSEC.WISC.EDU
<LOCAL-DATA> indicates that data will be accessed from the local data
directory.DATALOC -- done
So, even though your XCD decoder was writing to your local MD files, your
PTLIST (ptlist.k) invocations were trying to access datasets on a remote
machine.
I changed the pointing for RTPTSRC to LOCAL-DATA, and then ran PTLIST
looking for MLRF1 obs, and was successful:
mcidas* dataloc.k ADD RTPTSRC LOCAL-DATA
Group Name Server IP Address
-------------------- ----------------------------------------
RTPTSRC <LOCAL-DATA>
<LOCAL-DATA> indicates that data will be accessed from the local data directory.
mcidas* ptlist.k RTPTSRC/SHIPBUOY.6 SEL='TYPE CMAN;IDA MLRF'
Row : 1 Col : 114
DAY = 2003036 CYD | TIME = 0 HMS |
CMAX = 1663 | MOD = 0 |
TYPE = CMAN | IDA = MLRF |
IDB = 1 | HMS = 0 HMS |
LAT = 25.0100 DEG | LON = 80.3800 DEG |
CA = _missing_ | T = 295.66 K |
TD = _missing_ | STI = 295.26 K |
DIR = 280 DEG | SPD = 3.6 MPS |
PSL = 1015.7 MB | VIS = _missing_ |
WXP = _missing_ | SWP = _missing_ |
SWH = _missing_ | PSD = _missing_ |
PSP = _missing_ | PSH = _missing_ |
PCP = _missing_ | PTIM = _missing_ |
SP10 = 3.6 MPS | SP20 = 3.6 MPS |
SPMX = 4.6 MPS | LVLW = _missing_ |
---------------------------------------------------------------------------
Number of matches found = 1
ptlist.k: Done
The last listing demonstrates an important point. McIDAS stores the 5
character CMAN IDs in two keys: IDA and IDB. If you are going to search
for MLRF1, you need to either specify IDA= and IDB= in your PTLIST SEL=
clause, or, since there is no MLRF2, MLRF3, etc. only search for
MLRF:
% ./cman_dumper0.sh MLRF 0
Since you are running XCD, you should change the pointing for all
datasets that will be populated with XCD-created data files. The
main set is RTPTSRC, and I have just changed this for you. The
second is RTWXTEXT. Right now, you are pointing at a machine
at the University of Northern Illinois for that:
...
RTWXTEXT WEATHER2.ADMIN.NIU.EDU
...
I would change this to be local access with:
<as 'mcidas'>
cd ~mcidas/workdata
dataloc.k ADD RTWXTEXT LOCAL-DATA
Now, if you are every going to access the McIDAS datasets being created
locally on sapodilla from accounts other than 'mcidas' or 'ldm', or if
you can foresee ever accessing that data from machines at RSMAS other
than sapodilla, I strongly recommend that you setup the ADDE remote
server interface on sapodilla. This is quite easy, but it has more
than one step:
1) create the properly configured .mcenv file in ~mcidas
2) create a user named 'mcadde'. This user should not have a login
shell; should be in the same group as 'mcidas' and 'ldm'; and have
the same HOME directory as 'mcidas'
3) as 'root' install the ADDE remote server stuff in /etc/services
and /etc/inetd.conf:
<as 'root'>
cd ~mcidas
sh ./mcinet2002.sh install mcadde
If all three steps are done correctly, then you and other machines that
have internet access to sapodilla could point to your own machine for
all of the datasets that it hosts (e.g., RTIMAGES, CIMSS, RTPTSRC,
RTWXTEXT).
>Thanks,
On another topic, I did the following two things on sapodilla:
1) installed a McIDAS addendum that fixed a problem with the Synoptic/
Ship/Buoy decoder. The problem was that the decoder would use
excessive amounts of CPU on certain operating systems (RedHat 8
and Compaq OSF/1 being the two where I have seen the problem).
The addendum also fixed a number of other bugs and made various
enhancements. I made this change last Saturday
2) installed a pre-release version of LDM 6, and this appears to be running
well on your machine. I made the upgrade in preparation for transferring
feed responsibilities for the University of Rio de Janerio from
UCAR to you (as we talked about recently). I made the LDM change
yesterday.
Please let me know if you see anything out of the expected on sapodilla
both in LDM and McIDAS land.
Tom