[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
20031121: 20031121: cpu usage by dcnexr2
- Subject: 20031121: 20031121: cpu usage by dcnexr2
- Date: Fri, 21 Nov 2003 14:24:02 -0700
Eric,
I'd probably suggest grabbing 5.6.L.1 (at least dcnexr2 from the binary...
which may in fact be what I gave you back when).
The pqact.gempak_craft entry I provide in $NAWIPS/ldm/etc/templates looks like:
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 /dev/null data/craft/\1/\1_\2_\3
The (....) will provide a different \1 matching for each site, so you
will have a separate dcnexr2 for each site.
I run 3 pqact processes from ldmd.conf:
exec "pqact -f CRAFT -p BZIP2/K[A-E]
/local/ldm/etc/GEMPAK/pqact.gempak_craft"
exec "pqact -f CRAFT -p BZIP2/K[F-L]
/local/ldm/etc/GEMPAK/pqact.gempak_craft"
exec "pqact -f CRAFT -p BZIP2/K[M-Z]
/local/ldm/etc/GEMPAK/pqact.gempak_craft"
This way, each of the pqact processes will have less than 32 open file
descriptors
and balance out the file open/closes needed on the system.
Steve Chiswell
>From: Eric Nelson <address@hidden>
>Organization: UCAR/Unidata
>Keywords: 200311212041.hALKfaEH015256
>
>Steve,
>
>I beleive we are using a special version of dcnexr2. the dcnexr2 from
>ver "k" did not like our system. You gave us a compiled
>replacement. Also, should there be a seperate instance of dcnexr2 for
>each site we request?
>
>dcnexr2 is always the top gempak decoder in terms of the cpu usage on our
>system.
>
>-Eric
>
>On Fri, 21 Nov 2003, Unidata Support wrote:
>
>>
>> Eric,
>>
>> What version are you running?
>>
>> In general, the process should be short lived- but bunzip isn't exactly chea
> p.
>> I have all 61+ sites comming in here and being decoded.
>>
>> The CPU% output is dependent on what else you have coming in, eg, if
>> nothing else is being decoded, then virtually all cycles will go
>> to that 1 process. But, if the decoder is hung, then
>> ot would use lots of CPU and the TIME would appear large.
>>
>> Steve Chiswell
>>
>>
>>
>>
>> >From: Eric Nelson <address@hidden>
>> >Organization: UCAR/Unidata
>> >Keywords: 200311211903.hALJ3CEH010403
>>
>> >
>> >I was wondering what the typical cpu usage should be for dcnexr2. When I
>> >run top, dcnexr2 is almost always the "top" of the list for cpu
>> >usage. It's not uncommon for me to see....
>> >
>> > PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
>> >21677 ldm 18 0 924 924 504 S 99.9 0.0 3:54 dcnexr2
>> >
>> >As of right now I'm only ingesting one site. But if this is normal
>> >behavior, I'd be affraid to ingest too many more sites.
>> >
>> >Is this what I should be seeing?
>> >
>> >Thanks.
>> >-Eric
>> >
>> >--
>> >________________________________________________________________________
>> >Eric Nelson |
>> > |
>> > +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ |
>> >+ College of Dupage Meteorology Dept. + Research Applications Prog. + |
>> >+ Weather Data Manager + UCAR/NCAR + |
>> >+ http://weather.cod.edu + http://www.rap.ucar.edu + |
>> >+ address@hidden + address@hidden + |
>> > +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ |
>> > |
>> >"I love deadlines. I especially like the whooshing sound they make as |
>> >they go flying by." --Douglas Adams |
>> >________________________________________________________________________
>> >
>>
>> ****************************************************************************
>> 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
>> ****************************************************************************
>>
>