Week Eight
July 18th - July 24th, 2011
This weekend, we took a trolley ride through Fort Collins! Then we visited the local Avery House Museum that belonged to a prominent Fort Collins resident about a hundred years ago. He used to be the City Planner and thanks to him, Fort Collins has very spacious roads that allow for tons of wide bicycle lanes everywhere.
The week started with us trying to release BGPmon. This was
a failure. It turns out that our testing environment didn't
exactly match the high volume of input that the production
server of BGPmon handles. Several startup bugs and possibly
corrupt MRT messages were identified. For now, we don't know
a definite public release of version 7.2.1 code. Another
problem we uncovered was that an in-development alpha version
of BGPmon 7.2 has been on the public website since January!
So now we have to call this version "7.2.1" even though it's
technically 7.2. Oh well, just a small numbering error.
This week, Cathie shows me various debugging command-line
tools for both bugs and memory leaks. The first was gdb
for BGPmon code, followed by the perl debugger for BGPmon
web client code, and finally valgrind for memory leaks. Since
the web client that displays BGPmon real-time data is finally
running, the priority is now to help Jason and Cathie with
bugs and memory leaks. We suspect the large memory footprint
is due to memory leaks and that these leaks also have been
causing mysterious 4am crashes on the main BGPmon box.
For now, we have to explain about 5 possibly corrupt MRT
messages in our test data input stream.
|
