Ohhhh..the fun continues. I have been troubleshooting offline with Daryl.
So, from what I can tell, the level2 files for some radars (for whatever
reason) are transmitting two volume scans per file. The radars in question
appear random, but it might be related to a recent upgrade in ROC software or
something. EDIT: the files that blow things up are SAILS_1!!!
The file sizes are double because (surprise!) there are two timesteps due to
this SAILS_1.
Copy of my pqact here, but I assume this is going to create somewhat of a
headache for others as well.
https://atlas.niu.edu/vgensini/temp/pqact.split.radar
-----------------------------------------------
Vittorio (Victor) A. Gensini, Ph.D., CCM
Associate Professor
Department of Geographic and Atmospheric Sciences
Northern Illinois University
DeKalb, IL 60115
https://atlas.niu.edu<http://atlas.niu.edu>
https://wcs.niu.edu<http://wcs.niu.edu/>
<http://a><http://atlas.niu.edu>
________________________________
From: Herzmann, Daryl E [AGRON] <akrherz@xxxxxxxxxxx>
Sent: Tuesday, September 29, 2020 11:38 AM
To: Victor Gensini <vgensini@xxxxxxx>; ldm-users@xxxxxxxxxxxxxxxx
<ldm-users@xxxxxxxxxxxxxxxx>
Subject: Re: level2 via LDM
Well Howdy,
This sounds like a queue that is not holding an hour's worth of data and thus
getting duplicated / triplicated product transmisions. My questions:
1) Is your LDM queue holding an hour's worth of data `pqmon`? Do you know if
your upstream sources are holding an hour's worth of data? Is your LDM logging
anything exciting about switching upstreams etc ?
2) Do the file write timestamps on the files make sense? Such that the write
time coincides with the end of the VCP transmission?
3) Do your file byte sizes match what is found on the files at AWS / my
realtime site?
https://mesonet-nexrad.agron.iastate.edu/level2/raw/KDVN/
Do you have a recent example (today / yesterday) of a 2x / 3x file? Can you
share that file for others to interrogate!
daryl
--
/**
* daryl herzmann
* Systems Analyst III -- Iowa Environmental Mesonet
* https://mesonet.agron.iastate.edu
*/
________________________________________
From: ldm-users <ldm-users-bounces@xxxxxxxxxxxxxxxx> on behalf of Victor
Gensini <vgensini@xxxxxxx>
Sent: Tuesday, September 29, 2020 11:30 AM
To: ldm-users@xxxxxxxxxxxxxxxx
Subject: [ldm-users] level2 via LDM
Tom/Steve and LDM junkies,
Have there been any recent changes to the format of LEVEL2 radar files coming
over LDM that I perhaps missed?
A couple of weeks ago, I had a couple scripts start to break that are using
gpnexr2_gf with:
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
#0 0x7FD1DF3C1697
#1 0x7FD1DF3C1CDE
#2 0x7FD1DE8BC3AF
#3 0x7FD1DE9F0B54
#4 0x4BCFCB
#5 0x4B8E59
#6 0x4B93C7
#7 0x42A510
#8 0x4288FE
#9 0x40F5A7
#10 0x40903D
#11 0x40554A
#12 0x405588
#13 0x7FD1DE8A8504
#14 0x4046D7
#15 0xFFFFFFFFFFFFFFFF
Segmentation fault (core dumped)
A quick peek of the file size for these stations (relative to others) indicate
nearly a doubling (in some cases tripling) of the file size versus other
stations that are still working. As an example, compare your file sizes today
for KDVN and KILX, both of which are in VCP 35.
I'm using the standard hhmmssRadarII perl script to create the volumes off of
LDM. Again, no issues for some sites, and then this memory error is popping up
for others. I sense a change in file contents or format...
Any/all ideas welcome.
Victor
-----------------------------------------------
Vittorio (Victor) A. Gensini, Ph.D., CCM
Associate Professor
Department of Geographic and Atmospheric Sciences
Northern Illinois University
DeKalb, IL 60115
https://atlas.niu.edu<http://atlas.niu.edu>
https://wcs.niu.edu<http://wcs.niu.edu/>
<http://a><http://atlas.niu.edu>
