27 Sep 2000 jbowman   » (Journeyer)

Hrm. Installed RedHat 7.0 last night. Found it refuses to upgrade 7.0 beta installs due to some strange python segfault in the installer. I'd report it as a bug, but I know the response I'd get: "We don't support or recommend upgrading a beta release to the released version". Feh

jlf continues to poke and prod at me in an attempt to nudge me around some of the obstacles he's encountered in the industry. Thanks for being a mentor, Jeremy. In discussing my last diary entry with him, I think I may have described things a bit inaccurately. The "waiting game" I grow tired of isn't the wait time caused by waiting on customers to get answers back, but rather the wait time in finding out where the heck our end of things stands. It's the complete lack of information being handed out that irks me, not the actual act of waiting itself. To analogize it to forked processes, it's the difference between being a zombie child process that gets reaped and a zombie child process that doesn't and just sits around all day...

ahosey and I have been discussing other uses for spinwebd/logwatchd (I really do need to change the name soon). What we want/need at work in the not-to-distant future is a log multiplexor.
For example:
We have several thousand customers hosted on our webserver.
We want to be able to sort their webpage statistics out for them.
We can't do this because, well, the log traffic for that number of users is absolutely obscene.
So, what we can do with logwatchd is setup the appropriate regex and have it append the contents of the line it matched to a specific file based on the regex. This would give us the ability to record all of user foo's log entires into file bar on the system. And the nice thing is that unlike a bulky log-parsing/sorting script, this works "live" as the data comes in.

After saying that logwatchd could already do that with very little in the way of changes, Adrian pointed out that with as much traffic as is being generated, that's an awful waste of cpu cycles. We got into a discussion of LRU caching of open file descriptors, and how best to organize things. Now I just need to get my workstation at home up and working again, and I'll be able to add a new module to logwatchd.

Adrian also pointed out that my feelings of "Surely someone's done this before!" aren't all that unusual. His explanation lay (basically) along the lines of "Lots of people have ideas, few of them actually get implemented, and even fewer are ever shown to the world".I guess that's what I like the most about the spirit of open source software. You get the chance to reuse other peoples ideas (and, assuming the license is compatible, their code as well) to inspire new solutions to problems.

[ "We are the all-singing, all-dancing crap of the world." ]

Latest blog entries     Older blog entries

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!