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

20031120: Level II display



Tony,

The dcnexr2 program will initially open the file preceeded by a "."
character, and on closing the file will be renamed. This is done
so that files will be found containing a full scan of data,
rather than possibly a single 100 radial chunk (which makes
incomplete pictures in GUIs that auto update - like NMAP2).

In this case, you probably don't have a lot of open file descriptors,
so the file isn't closed until a timeout occurs. If you don't have
many sites requested, you might want to change the PIPE to "PIPE<tab>-close"
so that you don't have to wait 8 minutes. With more sites, or a busier
LDM with lots of decoders, you probably will see this lag go away.

The act of un-bzipping will be cpu intensive, but the CPU percentage should
only be representative of a system not doing much else. When you have
a lot of other decoders running, the CPU will be shared among them.
You may be better off looking at the load average.

Steve Chiswell

 
>From: address@hidden
>Organization: UCAR/Unidata
>Keywords: 200311202332.hAKNWwEH027937

>Don, 
>
>Sorry the text comes out funky. You're the first one to ever comment on
>it, so I didn't know it was a problem. I'm trying something different
>now that may make it worse or better. Let me know. The school went to a
>new mail system at the start of the semester (iPlanet), and to a person,
>we all hate it. Maybe I should try something else?
>
>Back to weather, I'll try reloading it on the Linux machine and see what
>happens. On the Windows side, we will have the ability to use Samba on
>the new server, so I'll have Kurt set that up. In the meantime, I've
>ftp'd a few Level II files over to my Windows laptop and I'm looking at
>them with 1.1b2 as we speak. They are awesome! I only wish we had some
>precip in the range of KFTG to make it even more interesting. Maybe this
>weekend, eh? The autorotate RHI is terrific.
>
>The LDM ingest and the decoding of the Level II data is another story.
>Seems each file is made up of several files that come in over a period
>of several minutes and the decoding can take up to 8 minutes with cpu
>usage ranging from 50-99% during that time! Wow. Is this right, or do I
>have something wrong in my pqact.conf? Can't imagine what would happen
>if I wanted to bring in more stations.
>
>########################################################################## 
># Level II data from CRAFT 
>########################################################################## 
># 
>CRAFT BZIP2/(....)/(........)(....)(..) 
>PIPE decoders/dcnexr2 -s \l -d data/gempak/logs/dcnexr2.log 
>/export/data/ldm/gempak/nexrad/CRAFT/\1/\1-\2-\3 
>
>We'll be installing the new workstations in the lab over the break and
>they will all be SuSE machines which is why I'm anxious to make sure the
>IDV works properly on it.
>
>Thanks for the help and I'll keep you posted. 
>
>Tony 
>____________________________________________________ 
>
>Anthony A. Rockwood 
>Metropolitan State College of Denver 
>Meteorology Program 
>P.O. Box 173362, Campus Box 22 
>Denver, CO 80217-3362 
>
>Office: 303-556-8399 
>Fax: 303-556-4436 
>Website: clem.mscd.edu/~rockwooa
>