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

20030829: 20030828: 20030828: 20030828: 20030828: 20030827: 20030827: 20030827: 20030826: dcnexr2 write errors



Eric,

Your glibc has non-standard defaults for processing the command line 
arguments. I have a work around. You you want the source or binary?

Steve Chiswell



>From: Eric Nelson <address@hidden>
>Organization: UCAR/Unidata
>Keywords: 200308300106.h7U16mLd015885

>
>Steve,
>
>I broke up the regular expresssion as suggested.  ldmpqactcheck was happy
>again.  I HUP'd pqact and....still not working right. 
>
>I went back to the previous email with instructions to test the decoder
>outside of ldm.  Used pqcat to make a file with all the pieces as
>suggested.  This yeilded a 676k test.klot file.  No error noted onscreen.
>
>I tried to run this file thru the decoder using...
>cat test.klot | dcnexr2 -s KLOT -v 4 -d - test.out
>
>This produced 2 files.  One file was named "4" and was 9.3M and the second
>file was KLOT and was 269 bytes.  "4" appeared to be the data while KLOT
>ended up looking like a log file.  Contents of KLOT appear below...
>
>[23938] 030829/2000 [DC 3]  Starting up. Version 5.6.k
>[23938] 030829/2000 [DCNEXR2 -1] new output file .4
>[23938] 030829/2000 [DC 5]  Normal termination.
>[23938] 030829/2000 [DC 2]  Number of bulletins read and processed: 44
>[23938] 030829/2000 [DC 6]  Shutting down.
>
>It looks as if the decoder is the problem after all.
>
>-Eric
>
>
>
>On Fri, 29 Aug 2003, Unidata Support wrote:
>
>> 
>> Eric,
>> 
>> OK....now we may be looking at a pqact parsing problem with pqact.conf.
>> To shorten the regular expression, I'd suggest:
>> CRAFT        BZIP2/(....)/(........)(....)
>>      PIPE    decoders/dcnexr2 -s \1 
>>              -d data/gempak/logs/dcnexr2.log
>>              data/gempak/craft/\1/\1_\2_\3
>> 
>> The regular expression above is sortened, and you can use multiple
>> lines for the action (using TABS for the continuation lines of course).
>> Just for the sake of clarity, the above is:
>> CRAFT <TAB> BZIP2/(....)/(........)(....)
>>       <TAB> PIPE <TAB> decoders/dcnexr2 -s \1
>>       <TAB> <TAB> -d data/gempak/logs/dcnexr2.log
>>       <TAB> <TAB> data/gempak/craft/\1/\1_\2_\3
>> 
>> 
>> Steve Chiswell
>> 
>> 
>> 
>> 
>> >From: Eric Nelson <address@hidden>
>> >Organization: UCAR/Unidata
>> >Keywords: 200308290239.h7T2dnLd006481
>> 
>> >
>> >Steve,
>> >
>> >Reread pqact (pqactHUP) and the files are still being written into the
>> >places I mentioned before.  I then tried specifying an explicit path of
>> >/home/ldm/data/gempak/craft/ with no luck.  It seems taht I could place
>> >just about anything in the pqact entry and it would make no
>> >difference.  .KLOT still writes to /home/ldm/ and KLOT is created as a
>> >file and not a directory.  
>> >
>> >Something else I noticed, I've copied and pasted the default entry
>> >different places with in my pqact file so I can test different paths after
>> >"PIPE". Each time I copy&paste, ldmadmin
>> >pqactcheck flags the pattern line for being too long, even though it is an
>> >exact copy of what I have a few lines above (which does not get flagged.)
>> >
>> >FYI we are running ldm 6.0.13 and debian linux 3.0
>> >
>> >-Eric
>> >
>> >On Thu, 28 Aug 2003, Unidata Support wrote:
>> >
>> >> 
>> >> Eric,
>> >> 
>> >> The GEMPAK decoders will write to the ~ldm/logs/ directory if
>> >> the -d path/name option isn't specified.
>> >> 
>> >> Presumably you have run "ldmadmin pqactHUP" or restarted the
>> >> LDM after making any edits to pqact.conf....since until
>> >> then, no changes you make will take effect.
>> >> 
>> >> Certainly go ahead and delete those files and make sure your pqact
>> >> has reread the entries. See if the files are recreated.
>> >> 
>> >> Steve Chiswell
>> >> 
>> >> 
>> >> 
>> >> 
>> >> >From: Eric Nelson <address@hidden>
>> >> >Organization: UCAR/Unidata
>> >> >Keywords: 200308282113.h7SLDnLd025401
>> >> 
>> >> >
>> >> >
>> >> >I found that dcnxer2.log is really being written in /home/ldm/logs/  and
>> >> >that a single file (not dir) named KLOT existed in /home/ldm/
>> >> >
>> >> >The log file is absent any errors. Standard start/stop stuff.
>> >> >
>> >> >[26238] 030828/1504 [DC 3]  Starting up. Version 5.6.k
>> >> >[26238] 030828/1520 [DC 5]  Normal termination.
>> >> >[26238] 030828/1520 [DC 2]  Number of bulletins read and processed: 28
>> >> >[26238] 030828/1520 [DC 6]  Shutting down.
>> >> >
>> >> >My pqact entry is still...
>> >> >
>> >> >PIPE     decoders/dcnexr2 -s \1 -d data/gempak/logs/dcnexr2.log data/gem
>> > pak/craf
>> >> > t/\1/\1_\2_\3
>> >> >
>> >> >so I'm puzzeled as to why it wants to write elsewhere.  
>> > >
>> >> >-Eric
>> >> >
>> >> >On Thu, 28 Aug 2003, Unidata Support wrote:
>> >> >
>> >> >> 
>> >> >> Eric,
>> >> >> 
>> >> >> I'd suggest storing off a copy of data from the LDM to disk and runnin
> g t
>> > he
>> >> >> decoder by hand to see if that provides a better handle.
>> >> >> 
>> >> >> You could do this like (use whatever pattern to get all pieces of a si
> ngl
>> > e
>> >> >> time:
>> >> >> pqcat -o 3600 -f CRAFT -p "20030828194338" -vl- > test.klot
>> >> >> 
>> >> >> Then pass the test.klot file to the decoder with:
>> >> >> 
>> >> >> cat test.klot | dcnexr2 -s KLOT -v 4 -d - test.out
>> >> >> 
>> >> >> It seems that if you aren't seeing a bunch of these decoders piling up
>> >> >> in you process list, and the ldmd.log isn't giving you any error broke
> n
>> >> >> pipe messages, then its possible that the data is being written
>> >> >> somewhere unexpected on your system.
>> >> >> 
>> >> >> Steve Chiswell
>> >> >> 
>> >> >> 
>> >> >> >From: Eric Nelson <address@hidden>
>> >> >> >Organization: UCAR/Unidata
>> >> >> >Keywords: 200308281929.h7SJTtLd013978
>> >> >> 
>> >> >> >
>> >> >> >Steve,
>> >> >> >
>> >> >> >climate:/home/data/gempak/craft> ls -la
>> >> >> >total 8
>> >> >> >drwxrwxrwx    2 ldm      ldm          4096 Aug 27 23:10 .
>> >> >> >drwxrwxrwx   18 ldm      ldm          4096 Aug 26 20:54 ..
>> >> >> >
>> >> >> >It has never created the KLOT dir (KLOT is the only site I'm requesti
> ng
>> >> >> >for now.)  I'd show you the log file its creating except it isn't cre
> ati
>> > ng
>> >> >> >one.
>> >> >> >
>> >> >> >right now I see a process...
>> >> >> > 2471 ?        R      2:37 decoders/dcnexr2 -s KLOT -d
>> >> >> >data/gempak/logs/dcnexr2.log data/gempak/craft/KLOT/KLOT_20030828_190
> 4
>> >> >> >
>> >> >> >but data/gempak/craft/ remains completely empty.
>> >> > >
>> >> >> >-Eric
>> >> >> >
>> >> >> >
>> >> >> >
>> >> >> >On Thu, 28 Aug 2003, Unidata Support wrote:
>> >> >> >
>> >> >> >> 
>> >> >> >> Eric,
>> >> >> >> 
>> >> >> >> If you have an instance of dcnexr2 running, it will create the dire
> cto
>> > ry
>> >> >> >> (KLOT presumably). Has this been done? The data file will initially
>  
>> >> >> >> be opened in the directory with a "." preceeding the filename speci
> fie
>> > d
>> >> >> >> in order to help with the fact that the data are sent 100 radials a
> t a
>> >  ti
>> >> > me.
>> >> >> >> The file will be renamed from .filename to filename the first time 
> the
>> >  de
>> >> > cod
>> >> >> > er 
>> >> >> >> exits.
>> >> >> >> 
>> >> >> >> Steve Chiswell
>> >> >> >> 
>> >> >> >> 
>> >> >> >> >From: Eric Nelson <address@hidden>
>> >> >> >> >Organization: UCAR/Unidata
>> >> >> >> >Keywords: 200308281400.h7SE0vLd017431
>> >> >> >> 
>> >> >> >> >
>> >> >> >> >Ok,
>> >> >> >> >
>> >> >> >> >I've recovered from my earlier typos in pqact so that everything i
> s
>> >> >> >> >working that was working with version J.  I copied all the dc* int
> o
>> >> >> >> >/home/ldm/decoders where they should be.  I'm not seeing anymore w
> rit
>> > e
>> >> >> >> >errors in my ldmd.log file.
>> >> >> >> >
>> >> >> >> >dcnexr2 is still not creating any output.  I used the sample entry
>  an
>> > d
>> >> >> >> >modified it to generate a log file in the same place the rest of m
> y l
>> > og
>> >> >> >> >files are.  I also changed the data output location for my data tr
> ee.
>> >   T
>> >> > he
>> >> >> >> >pattern is the same.
>> >> >> >> >
>> >> >> >> >CRAFT
>> >> >> >> >^L2-BZIP2/(....)/([0-9][0-9][0-9][0-9][0-1][0-9][0-3][0-9])([0-2][
> 0-9
>> > ][0
>> >> > -5]
>> >> >> > [0-
>> >> >> >> > 9])([0-9][0-9])
>> >> >> >> >   PIPE    decoders/dcnexr2 -s \1 -d data/gempak/logs/dcnexr2.log
>> >> >> >> >data/gempak/craft/\1/\1_\2_\3
>> >> >> >> >
>> >> >> >> >No log files, No data.  When I look at the list of processes runni
> ng 
>> > I s
>> >> > ee
>> >> >> >> >an instance of dcnexr2 running.
>> >> >> >> >
>> >> >> >> >The dir that it should be writing to has 777 permissions on it.
>> >> >> >> >
>> >> >> >
>> >> >> 
>> >> >> **********************************************************************
> ***
>> > ***
>> >> >> Unidata User Support                                    UCAR Unidata P
> rog
>> > ram
>> >> >> (303)497-8643                                                  P.O. Bo
> x 3
>> > 000
>> >> >> address@hidden                                   Boulder, CO
>  80
>> > 307
>> >> >> ----------------------------------------------------------------------
> ---
>> > ---
>> >> >> Unidata WWW Service              http://my.unidata.ucar.edu/content/su
> ppo
>> > rt 
>> >> >> **********************************************************************
> ***
>> > ***
>> >> >> 
>> > >
>> >> >
>> >> 
>> >> *************************************************************************
> ***
>> >> Unidata User Support                                    UCAR Unidata Prog
> ram
>> >> (303)497-8643                                                  P.O. Box 3
> 000
>> >> address@hidden                                   Boulder, CO 80
> 307
>> >> -------------------------------------------------------------------------
> ---
>> >> Unidata WWW Service              http://my.unidata.ucar.edu/content/suppo
> rt 
>> >> *************************************************************************
> ***
>> >> 
>> >
>> 
>> ****************************************************************************
>> Unidata User Support                                    UCAR Unidata Program
>> (303)497-8643                                                  P.O. Box 3000
>> address@hidden                                   Boulder, CO 80307
>> ----------------------------------------------------------------------------
>> Unidata WWW Service              http://my.unidata.ucar.edu/content/support 
>> ****************************************************************************
>> 
>