Over the weekend, I found Markmail via Farber's IP list, and they agreed to index our rbnb-dev mailing list. Check it out at http://markmail.org/search/?q=list%3Arbnb

Yesterday, Jeff Avila and I moved niagara and its RAID chassis from Atkinson to the SDSC machine room. (This is part of the process where we outsource sysadmin to them, for which I am quite grateful and should result in better uptime for all).

No problems, now installed in a CAIDA isle, coordinates A1 or so. I took a few snaps of the room, other machines and such for your amusement.

Album is at http://dataturbine.org/content/niagara-moves-sdsc-machine-room

Just got first clean square wave off of our Campbell CS-505 sensor, using a USB DAQ device from the shelf in the lab. I need to create and solder a BNC connector, but I'm getting 1120 Hz, which seems about right.

Tomorrow - solder, see if I can get it running on cRIO. Also need to setup the thermistor as well, then the UV FireScout.

Fun stuff, more later.

PS First light is terminology from Fermilab and the accelerator there.

See build status at http://nmi-web.cs.wisc.edu/nmi/index.php?page=results%2Foverview&rows=20...

Applied for an account today.

Since my mini is flaky, for reasons I've had trouble isolating, I'm moving and updating docs and Doxygen from it to Apache2 on the Sun. That should be reliable!

New URLs: http://niagara.sdsc.edu/doxygen/ and http://niagara.sdsc.edu/Documentation

Today I stopped all data turbines on niagara, -dev and -stage and updated them with current code out of Subversion. We want to test the improved mirroring, and for that we need the newest code. Rather than run different versions, I chose to upgrade all three at once and keep them synchronized.

I plan to leave them running over the weekend, and start mirror testing Monday if all looks good.

I've also moved the SMER and world-clock feeds from my crashy Mini to a new iMac; hopefully that'll be stable.

One of the questions we're working on is how to monitor a network infrastructure that includes DataTurbine. A typical instance might include the following:

1. Network time server
2. Data source machine, typically a data acquisition system
3. Video or image source
4. DataTurbine
5. Client machines running RDV
6. Network infrastructure: Routers, DNS server, switches, etc

Goals:
*One frame every five minutes from SMER
*One frame every 15 minutes from http://www.cru.uea.ac.uk/~timo/sunclock.htm

Method:
*Run RBNB on the mini, using parent-child routing to niagara-stage.sdsc.edu
*Run Tomcat on the mini for Webdav support
*Mount webdav locally so that we can use curl/wget to write images to DataTurbine

Problems, notes and caveats:
*I'd prefer to run Tomcat on niagara-stage, but I haven't solved the problem of binding Tomcat to specific IP yet.

Today we solved a problem we've had with the data from our Vaisala WX510 weather station at the Santa Margerita Ecological Reserve. We'd periodically get zeros, spikes or just erroneous data in the RDV viewer and didn't understand why.

Background: The Vaisala is connected to the main serial port on the cRIO, and is programmed to send updates automatically at approximately one-second intervals. 9600N81, no flow control.

In the course of solving a LabVIEW/cRIO problem, we added syslog support to the code and found it very useful.

Syslog (See http://en.wikipedia.org/wiki/Syslog) is an old Unix standard for centralized logging. You simply send short UDP packets, in ASCII, to port 514 on the logging host, and they are saved to a persistent logfile. (See http://tools.ietf.org/html/rfc3164 for protocol details)

So all that the labview has to do is

- Open a UDP socket to port 514 on the loghost
- Send a string out the socket
- Close it

The string is like this

<10>{message}