climode-iss2: SOFTWARE Messages: 6 Entries..

Data manager computer froze up about 2 UT.  Not sure what caused it,
but did see a message that too many x windows were open.

Had to do a hard reboot to get it working again.
Also restarted POP to can serial line comms working again.

I checked on the tilt sensor data with the 'dsmctl dump' command, and the tilt
sensor was reading constant values:


1 iss2:/home/iss> dsmctl dump
FileSet::setDSMConfig: FileSet: /iss/ds/raw_tilt_sensor/tilt_raw_%Y%m%d_%H%M%S.dat
|--- date time -------| deltaT   bytes
2007 02 18 18:46:07.913 1171824367913       8   0.002747   0.002747
2007 02 18 18:46:08.013 100       8   0.002747   0.002747
2007 02 18 18:46:08.113 100       8   0.002747   0.002747
2007 02 18 18:46:08.221 107       8   0.002747   0.002747
2007 02 18 18:46:08.321 100       8   0.002747   0.002747
...

I haven't determined when this started.  To fix it, I had to restart the dsm
and cycle the power on the tilt sensor.  After unplugging the tilt sensor
power supply and plugging it back in, then restarting dsm, no data were
coming in, so I restarted dsm a second time.  Now it appears to be back
to normal.

7 iss2:/home/iss> dsmctl stop
Killing dsm...
8 iss2:/home/iss> dsmctl start
Starting dsm...
+ dsm /iss/src/ingest/tilt_sensor/tilt_sensor.xml
9 iss2:/home/iss> dsmctl dump
received signal Interrupt(2), si_signo=2, si_errno=0, si_code=128
IOException: inet:localhost.localdomain:40986: recv: Interrupted system call
10 iss2:/home/iss> dsmctl stop
Killing dsm...
11 iss2:/home/iss> dsmctl start
Starting dsm...
+ dsm /iss/src/ingest/tilt_sensor/tilt_sensor.xml
12 iss2:/home/iss> dsmctl dump
FileSet::setDSMConfig: FileSet: /iss/ds/raw_tilt_sensor/tilt_raw_%Y%m%d_%H%M%S.dat
|--- date time -------| deltaT   bytes
2007 02 18 18:48:24.733 1171824504733       8     0.7883      1.104
2007 02 18 18:48:24.821 087       8     0.7388       1.09
2007 02 18 18:48:24.929 108       8     0.7251      1.011
2007 02 18 18:48:25.017 088       8     0.8212     0.9998
...

I've added a step to the sounding processing on the ship end.  After each
sounding is converted to netcdf with Aspen, the ncks program is used to
create a decimated netcdf file by picking off every 10th sample.  The
reduced files are put into /iss/ds/class.decimated, and those files have
been added to the list of datasend products.

On the receiving end, the 'class' directory is a link to class.decimated,
and so the usual web plotting mechanisms plot the decimated soundings on
the web.

I also added tklog to the data transfers, on a tentative basis, since the
entire file gets transferred each time it changes, and the log could become
too large to justify the bandwidth.

The ISS2 workstation froze last night, so I had to reboot it.  Apparently
the reboot did not happen fast enough, because POP stopped sending moments
over the serial connection.  I just noticed that now, so I've restarted
POP and the serial moments are ingesting again.

I've created a new script module, /iss/etc/init.d/dsmctl, and tied it into the
normal ISS startup, so the NIDAS dsm should start automatically now and 
record the raw tilt sensor data to /iss/ds/raw_tilt_sensor.

To check that the tilt angles are being read, run this command:

 dsmctl dump

To see some statistics about the data stream, such as sampling rate and
number of samples, use this command:

 dsmctl stats

You can run dsmctl by itself to see its command-line options.

We are underway as of about 9:30 EST this morning.

The shipdat.py script is listening to the broadcast UDP ship navigation data
and generating the ship.dat file from required by POP.  It runs on the
profiler PC.

There was a bug in the conversion from km/hour to m/s, so for the first few
hours the speed is way off, and consequently the winds calculated by POP.
However, with the artificially high speed we were able to verify that the
antenna orientation is working correctly in POP, because the wind direction
corresponds well with the ship's heading.

The lat/lon were fixed also, as the decimal minutes in the NMEA GGA sentences
were not being converted to decimal degrees correctly.  So the fix for that
was also installed.

The current indication is that the ship navigation data and POP wind
calculations are working correctly.