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

[McIDAS #YSZ-272620]: need help to setup a McIDAS ADDE server



Hi Hsie,

re: slightly modify your testing

> Do as you suggested.  Make no difference.

OK.  I did not expect a change; I wanted to make sure, however,
to point out that one should always run McIDAS commands from his/her
$MCDATA directory.

re: Questions

> squall:[70]% cd /home/mcidas/linux
> squall:[71]% ls -l
> total 128
> drwxrwxr-x  2 mcidas unidata  4096 May  2  2006 admin
> drwxrwxr-x  2 mcidas unidata 12288 Jan 24 11:21 bin
> drwxrwxr-x  3 mcidas unidata 20480 Jan 24 14:10 data
> drwxrwxr-x  2 mcidas unidata  4096 May  2  2006 etc
> drwxrwxr-x  2 mcidas unidata 12288 May  2  2006 help
> drwxrwxr-x  2 mcidas unidata  4096 May  2  2006 inc
> drwxrwxr-x  2 mcidas unidata  4096 May  2  2006 include
> drwxrwxr-x  3 mcidas unidata  4096 May  2  2006 lib
> drwxrwxr-x  4 mcidas unidata  4096 May  2  2006 man
> drwxrwxr-x  2 mcidas unidata  4096 May  2  2006 savedata
> drwxrwxr-x  6 mcidas unidata  4096 May  2  2006 tcl
> drwxrwxr-x  2 mcidas unidata  4096 Jan 25 01:20 workdata
> squall:[72]% cd /home/hsie
> squall:[73]% ls -ld mcidas
> drwxr-xr-x  3 hsie al4 4096 Dec  3 09:17 mcidas
> squall:[74]% cd mcidas
> squall:[75]% ls -l
> total 8
> drwxr-xr-x  2 hsie al4 4096 Jan 25 12:45 data
> squall:[76]%

This all look correct.

re: environment variables needed for McIDAS

> MCHOME=/home/mcidas/linux
> MCDATA=/home/hsie/mcidas/data
> MCPATH=/home/hsie/mcidas/data:/home/mcidas/linux/data:/home/mcidas/linux/help
> MCGUI=/home/mcidas/linux/bin
> MCTABLE_READ=/home/hsie/mcidas/data/MCTABLE.TXT;/home/mcidas/linux/data/ADDESITE.TXT
> MCTABLE_WRITE=/home/hsie/mcidas/data/MCTABLE.TXT
> MCCOMPRESS=TRUE

The only change I suggest here is to change MCCOMPRESS from TRUE to GZIP.  This
is optional since your machine has 'compress' and 'uncompress' installed and
both are accessible to the 'mcidas' user through its defined PATH.

re: recommendation to run current McIDAS release

> I had been using this machine for last summer's field experiment.  All
> the softwares I need are already installed in this machine (LDM,
> McIDAS-X, GEMPAK).  The 2005 version will serve my purpose.

Very good.

I logged onto squall this morning to snoop around.  Here is what I found:

- as you noted in your original email, I can run DATALOC (dataloc.k) with
  no problems as 'mcidas'

- since I couldn't become the user 'hsie' to see what may be going on in
  that account, I became 'ldm' and ran some tests there:

  <login as 'ldm'>
  cd /home/ldm
  mkdir mcidas
  mkdir mcidas/data
  cd /home/ldm/mcidas/data

squall:[59]% dataloc.k LIST

Group Name                    Server IP Address
--------------------         ----------------------------------------
AMRC                         UWAMRC.SSEC.WISC.EDU
BLIZZARD                     ADDE.UCAR.EDU
CIMSS                        SQUALL.AL.NOAA.GOV
GINICOMP                     SQUALL.AL.NOAA.GOV
GINIEAST                     SQUALL.AL.NOAA.GOV
GINIWEST                     SQUALL.AL.NOAA.GOV
GOESEAST                     UNIDATA2.SSEC.WISC.EDU
ME7                          IO.SCA.UQAM.CA
MYDATA                       <LOCAL-DATA>
NEXRCOMP                     SQUALL.AL.NOAA.GOV
PUB                          GP16.SSD.NESDIS.NOAA.GOV
RTGINI                       <LOCAL-DATA>
RTGRIDS                      SQUALL.AL.NOAA.GOV
RTIMAGES                     SQUALL.AL.NOAA.GOV
RTNEXRAD                     SQUALL.AL.NOAA.GOV
RTPTSRC                      SQUALL.AL.NOAA.GOV
RTWXTEXT                     SQUALL.AL.NOAA.GOV
TOPO                         <LOCAL-DATA>

<LOCAL-DATA> indicates that data will be accessed from the local data directory.
DATALOC -- done

As you can see, a DATALOC command run as 'ldm' works as expected.  This
tells me that the problem you encountered trying to run McIDAS as 'hsie'
is related to some setting in the 'hsie' account, or that you found and
fixed the problem after sending me your reply yesterday (did you find a 
problem?).

So, without access to the 'hsie' account, I can proceed no further on
troubleshooting the problem you reported.

Next:

While logged into squall as 'ldm', I decided to see if I could access
data in the various datasets that are defined.  These attempts met
with varying success.  Here is one example of a failure:

<logged in as 'ldm'>
setenv MCPATH 
/home/ldm/mcidas/data:/home/mcidas/linux/data:/home/mcidas/linux/help
cd /home/ldm/mcidas/data

squall:[60]% imglist.k RTIMAGES/GE-IR
Image file directory listing for:RTIMAGES/GE-IR
imglist.k: No images satisfy the selection criteria
imglist.k: done

The inability to list RTIMAGES images indicated that there was some problem
in the RTIMAGES dataset definition or there were no files to list.  The same
comment applies to GINIEAST, GINIWEST, GINICOMP, etc.  I was, however, able
to list out images from the NEXRCOMP/6KN0R-NAT dataset:

squall:[77]% imglist.k NEXRCOMP/6KN0R-NAT FORM=ALL
Image file directory listing for:NEXRCOMP/6KN0R-NAT
 Pos Satellite/         Date       Time      Center   Band(s)
     sensor                                 Lat  Lon
 --- -------------  ------------  --------  ---- ---- ------------
 301  RADAR         26 JAN 08026  16:34:00     TWX    1
imglist.k: done

A quick check shows that you are writing images to a different directory
structure than is specified for some of the ADDE datasets:

<as 'mcidas'>
cd $MCDATA
dsserve.k LIST GINIEAST
 ...
GINIEAST/GE4KIR          IMAGE GINI                  GINI 4 km 10.7 um East CONU
                                                     S
                               
DIRFILE=/wrk/data1/ldm/sat/gini/EAST-CONUS/4km/IR/IR
 ...

This entry, for example, tells McIDAS to look in the 
/wrk/data1/ldm/sat/gini/EAST-CONUS/4km/IR
directory for GOES-East IR (10.7 um) images from NOAAPORT (IDD feedtype NIMAGE).
The ~ldm/etc/pqact.conf action, on the other hand, is writing those images as 
follows:

# Zlib compressed NOAAPORT GOES-East/West GINI images
NIMAGE  ^satz/ch[0-9]/.*/(.*)/([12][0-9])([0-9][0-9])([01][0-9])([0-3][0-9]) 
([0-2][0-9])([0-5][0-9])/(.*)/(.*km)/
        PIPE    -close
        decoder/bin/zlibg2gini -vl logs/png.log
        data/sat/gini/\8/\9/\1/\1_\2\3\4\5_\6\7

The ~ldm/data directory is a link:

squall:[51]% ls -alt data
lrwxrwxrwx  1 ldm unidata 17 Jun 20  2007 data -> /export/data1/ldm

If you intend to keep the LDM decoding actions the same, you will need to
modify the ADDE dataset definitions in the 'mcidas' account.

Also, your ~ldm/etc/pqact.conf file does not have any action(s) to process
UNIWISC imagery (aka MCIDAS LDM feedtype), so access to the RTIMAGES
datasets will always result in messages like:

squall:[56]% imglist.k RTIMAGES/GE-IR
Image file directory listing for:RTIMAGES/GE-IR
imglist.k: No images satisfy the selection criteria
imglist.k: done

On the bright side, data from the RTPTSRC and RTGRIDS datasets are accessible
and look correct:

squall:[59]% ptlist.k RTPTSRC/PTSRCS FORM=FILE ALL
Pos      Description                        Schema  NRows NCols Proj# Created 
DataDate
------   --------------------------------   ------  ----- ----- ----- ------- 
--------
     4   SAO/METAR data for   24 JAN 2008   ISFC       72  7000     0 2008024 
2008024
     5   SAO/METAR data for   25 JAN 2008   ISFC       72  7000     0 2008024 
2008025
     6   SAO/METAR data for   26 JAN 2008   ISFC       72  7000     0 2008025 
2008026
    14   Mand. Level RAOB for 24 JAN 2008   IRAB        8  1500     0 2008024 
2008024
    15   Mand. Level RAOB for 25 JAN 2008   IRAB        8  1500     0 2008024 
2008025
    16   Mand. Level RAOB for 26 JAN 2008   IRAB        8  1500     0 2008025 
2008026
    24   Sig.  Level RAOB for 24 JAN 2008   IRSG       16  6000     0 2008024 
2008024
    25   Sig.  Level RAOB for 25 JAN 2008   IRSG       16  6000     0 2008024 
2008025
    26   Sig.  Level RAOB for 26 JAN 2008   IRSG       16  6000     0 2008025 
2008026
    34   SHIP/BUOY data for   24 JAN 2008   ISHP       24  2000     0 2008024 
2008024
    35   SHIP/BUOY data for   25 JAN 2008   ISHP       24  2000     0 2008025 
2008025
    36   SHIP/BUOY data for   26 JAN 2008   ISHP       24  2000     0 2008026 
2008026
    54   SYNOPTIC data for    24 JAN 2008   SYN         8  6500     0 2008024 
2008024
    55   SYNOPTIC data for    25 JAN 2008   SYN         8  6500     0 2008024 
2008025
    56   SYNOPTIC data for    26 JAN 2008   SYN         8  6500     0 2008025 
2008026
ptlist.k: Done

squall:[60]% dsinfo.k G RTGRIDS

        Dataset Names of Type: GRID  in Group: RTGRIDS

Name         NumPos   Content
------------ ------   --------------------------------------
ALL           1210    Real-Time Grids
AVN            200    Real-Time AVN Grids
ECMWF           10    Real-Time ECMWF Grids
ETA            120    Real-Time ETA Grids
GFS            200    Real-Time GFS/AVN Grids
MDR             10    Real-Time MDR/AVN Grids
MISC            10    Other Real-Time Grids
MRF            200    Real-Time MRF Grids
MRF-UW          10    MRF Grids in UW stream
NAM            120    Real-Time NAM Grids
NGM             40    Real-Time NGM Grids
NGM-UW          10    NGM Grids in UW stream
RUC             80    Real-Time RUC Grids
UKMT           110    Real-Time UK Met Grids

DSINFO -- done

squall:[61]% grdlist.k RTGRIDS/GFS
Dataset position 5      Directory Title= ALL  00Z GFS 0  HR<=VT<=24 HR
PAR    LEVEL       DAY        TIME    SRC FHOUR     FDAY       FTIME  GRID  PRO
---- --------- ------------ -------- ---- ----- ------------ -------- ----- ----
T     850 MB   25 JAN 08025 00:00:00  GFS    12 25 JAN 08025 12:00:00     1 MERC
Number of grids listed = 1
GRDLIST - done

Wrap-up:

- the McIDAS v2005d installation on squall is working correctly

- the ADDE remote server on squall is working correctly (I even tested this from
  the 'mcidas' account on cirrus)

- I can not explain why DATALOC (dataloc.k) hangs from the 'hsie' account; it
  works correctly from the 'mcidas' and 'ldm' accounts on squall.  I could
  troubleshoot further if I had login access as 'hsie' on squall.

- there is still work to be done to process UNIWISC imagery in the LDM if
  you actually want those images

- you will need to either adjust your LDM pqact.conf actions for NIMAGE data
  or adjust your ADDE dataset definitions if you want to access imagery
  in the GINIEAST, GINIWEST, and GINICOMP datasets

Cheers,

Tom
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
(303) 497-8642                                                 P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage                       http://www.unidata.ucar.edu
****************************************************************************


Ticket Details
===================
Ticket ID: YSZ-272620
Department: Support McIDAS
Priority: Normal
Status: Closed