The data from Eaglecrest for the 2008/9 season is apparently flowing down to nsrl--and being caught by the "smvaio" machine. Time to get plots, web sites, etc. Josh G has (with Marijke's help) gotten the data inserted into the seamonster database (just on the nsrl1.jun.alaska.edu/browser "development database" for now).
This data goes over a 900 MHz freewave modem to the NSRL and a modem into vaio (18.104.22.168). Home written software (~seamonster/eaglecrest is a perl script watching the serial port and doing the crudest of data quality control) logs the data to an ASCII file.
From here, the data is going over to the seamonster database. It was doing the below (and we'd like to recreate that)--it is currently at http://nsrl1.jun.alaska.edu/smvaio/eaglecrest/eaglecrest.php
On smvaio (22.214.171.124 right now), I found ~seamonster/eaglecrest_rrd/run.py this does insert and update rrd databases (in ~seamonster/eaglecrest_rrd/rrd) and generate plots (in both ~seamonster/eaglecrest_rrd/images/[daily,monthly,sixhour,twelvehour,weekly] and in /var/www/eaglecrest/images/graphs/eaglecrest/[daily,monthly,sixhour,twelvehour,weekly] Running this script does create proper files in ~seamonster/eaglecrest_rrd_data (specifically data.backup--if you want to try running it multiple times, make sure you rm data.backup if you want to import data!)
It appears that although the time stamp on the file updates, and new graphs are generated, no new data is actually being inserted. is this a formating issue? I think one change is the "empty line" between each lines. a simple perl "chop" used to be in the ~seamonster/eaglecrest equivalent to fix this. Oh--and the values (column layout) may probably have changed--perhaps this is it? Or maybe the error is in the insert to rrd database? maybe here:
def importData(rrd, timestamp, data): try: # Import into an rrd database. rrdtool.update(dataPath + '/rrd/' + rrd, timestamp + ':' + data) except: pass # If an error is encountered when importing, skip the data point. Should probably write something to an error log.....
Nathan Rogers was the architect (based on Ed Knuth's work when Kent and Dom were first getting the station going.) I think last year this was running on the old black Dell tower. The user "root" has a cronjob running /var/www/eaglecrest/update_programs/run.sh every five minutes and spitting any errors out to the file /var/www/eaglecrest/update_programs/error_log.cron