20030918: McIDAS: Data Plotting Problems

>From: Michael Keables <address@hidden>
>Organization: DU
>Keywords: 200309181656.h8IGuJLd005114 McIDAS SFCPLOT SFCCON

Hi Mike,

>I haven't been able to plot or contour surface data for some time (haven't
>had a chance to look at the problem until now.) When I try to plot station
>data or contour maps, there are no data available to plot.
>
>Suggestions?

I just pointed my McIDAS session at the ADDE server running on cyclone
and see that you have current surface and upper air data.  I do see a
problem with the NLDN lightning data files, however, but let's leave
that until later.

Here is what I did to check out what you have:

DATALOC ADD RTPTSRC CYCLONE.NATNET.DU.EDU
PTLIST RTPTSRC/PTSRCS FORM=FILE ALL
Pos      Description                        Schema  NRows NCols Proj# Created 
DataDate
------   --------------------------------   ------  ----- ----- ----- ------- 
--------
     1   SAO/METAR data for   18 SEP 2003   ISFC       72  6500     0 2003261 
2003261
    10   SAO/METAR data for   17 SEP 2003   ISFC       72  6500     0 2003260 
2003260
    11   Mand. Level RAOB for 18 SEP 2003   IRAB        8  1500     0 2003261 
2003261
    20   Mand. Level RAOB for 17 SEP 2003   IRAB        8  1500     0 2003260 
2003260
    21   Sig.  Level RAOB for 18 SEP 2003   IRSG       16  6000     0 2003261 
2003261
    30   Sig.  Level RAOB for 17 SEP 2003   IRSG       16  6000     0 2003260 
2003260
    31   SHIP/BUOY data for   18 SEP 2003   ISHP       24  2000     0 2003261 
2003261
    40   SHIP/BUOY data for   17 SEP 2003   ISHP       24  2000     0 2003260 
2003260
    41   NGM MOS for day      18 SEP 2003   FO14       38   600     0 2003261 
2003261
    50   NGM MOS for day      17 SEP 2003   FO14       38   600     0 2003260 
2003260
    51   SYNOPTIC data for    18 SEP 2003   SYN         8  6500     0 2003260 
2003261
    60   SYNOPTIC data for    17 SEP 2003   SYN         8  6500     0 2003259 
2003260
    61   PIREP/AIREP data for 18 SEP 2003   PIRP       24  1500     0 2003261 
2003261
    70   PIREP/AIREP data for 17 SEP 2003   PIRP       24  1500     0 2003260 
2003260
    71   ????????????????????????????????   ????    ***** ***** ***** ******* 
*******
    80   ????????????????????????????????   ????    ***** ***** ***** ******* 
*******
PTLIST: Done

The bogus stuff in the listing for MD files 71 and 80 show that there is
a problem with the Lightning data.  The listing for files 1 & 10, 11 & 20,
and 21 & 30 show that you have files of current surface and upper air
data.

SFCMG KDEN 24
SFCMG: done

This worked fine showing that the data at least for Denver has been
received and decoded.

SFCPLOT T NACONF
Accessing Dataset Name = RTPTSRC/SFCHOURLY.ALL
SFCPLOT - Done

This plotted a North American map in CONFormal projection with a complete
set of current temperature data.

The above tells me that there is nothing wrong with your data, so the
problem must be in the McIDAS plot/contour routines or, more likely, a
return of an old problem you have run into before: a full file system.

I then logged onto cyclone as 'mcidas' and I noticed that /var was full:

% df -k
Filesystem            kbytes    used   avail capacity  Mounted on
/proc                      0       0       0     0%    /proc
/dev/dsk/c0t0d0s0      96455   45683   41127    53%    /
/dev/dsk/c0t0d0s6     877790  604703  211642    75%    /usr
fd                         0       0       0     0%    /dev/fd
/dev/dsk/c0t0d0s1     413639  405565       0   100%    /var
/dev/dsk/c0t1d0s7    8509324 7646027  778204    91%    /data
/dev/dsk/c0t0d0s7    4031022 3564853  425859    90%    /export/home
/dev/dsk/c0t0d0s5    3007086 1466969 1479976    50%    /opt
swap                  615288   24112  591176     4%    /tmp

The next thing I did was to run the same surface plot command
as 'mcidas' on cyclone, and it worked:

<login to cyclone as 'mcidas'>
cd workdata
% sfcplot.k T NACONF
Accessing Dataset Name = RTPTSRC/SFCHOURLY.ALL
SFCPLOT - Done

A contour plot, however, fails:

% sfccon.k T OLAY FRAME
Accessing Dataset Name = RTPTSRC/SFCHOURLY.ALL
PTCON: No data found matching search conditions
PTCON:  Done
sfccon.k: PTCON command failed 1
SFCCON - Done


This strikes me as something I have seen before on cyclone, and it was then
caused by /var being full.  To test this out, I set the TMPDIR environment
variable to point at /tmp (it defaults to /var/tmp under Solaris),
and things worked:

% setenv TMPDIR /tmp
% sfccon.k T OLAY FRAME
Accessing Dataset Name = RTPTSRC/SFCHOURLY.ALL
Number of data points input to objective analysis:  2225
PTCON:  Done
SFCCON - Done

This proves that the problem is that /var is full.  The temporary workaround
is for each user to set the TMPDIR environment variable as I did above.

Since I do not have 'root' access to the machine, I am unable to do
needed cleanup.  If you provide me the password (send to
address@hidden without reference to it being root's password
and without reference to the machine it pertains to), I could get on
and perhaps clean up enough space to get things working again.

>Thanks in advance.

No worries.

Tom