[an error occurred while processing this directive] >>> Back to post index <<<

2009/07/19 Upgraded to Jaunty: Network Manager Still Sucks Balls, 2.6.30.1 not that stable, and Comcast Business Exceeding Expectations
π 2009-07-19 01:01 in Linux
I had been running a half upgraded Jaunty for a while now and it was time to upgrade to full Jaunty to 2.6.30.1 (including the recent security fix).

The upgrade was relatively eventless after I realized that pulseaudio had broken my sound and some X server problem broke mythfrontend (alias mythfrontend='XLIB_SKIP_ARGB_VISUALS="1" \mythfrontend made it all better).

And then, I figured I'd try network manager again, just to be fair.
Well, after close to an hour, I could not get it to manage any of my interfaces, even after commenting them out in /etc/network/interfaces and playing with its sparse badly documented config files.
I searched a net for a howto or something that would explain how to beat that piece of shit into submission, but it would not manage either my wired or wireless interfaces, and output bullshit messages in syslog like
NetworkManager: <info> Unmanaged Device found; state CONNECTED forced. (see http://bugs.launchpad.net/bugs/191889)

Its /usr/share/doc still had no useful documentation, so I got fed up quickly and just removed the piece of crap, and re-installed wicd which was up and running 20mn later.

Network Manager probably only works for wipe-reinstalls, which I'm not about to do, and is otherwise the most unintuitive, badly documented, and still buggy piece of software I've had the displeasure of working with. Seriously, how many years will it be before it doesn't suck balls for people who try to upgrade and may have their own network configuration already?

In other news, I upgraded both my laptop and server magic to 2.6.30.1, but it proved to be slightly unstable on my laptop and majorly unstable on my server, which forced me to drive to the colo twice to power cycle it (BUG spinlock lockup on CPU#1 nicely prevents serial console sysrq), and in the end I just drove the server home in the weekend to make sure it was still stable on the old kernel (it was), and then racked it back 2 days later. So far, so good, so the good news is that my hardware is likely still good and that the problem is triggered by 2.6.30.1 (vs 2.6.27.8) and/or some changed config options.

Another small but interesting fact: I got comcast business to actually do reverse DNS delegation for my block of static IPs from them. I did have to walk the guy through replacing the PTR records with CNAMEs pointing to my name server, but in the end, it was done 10mn later on a sunday morning at 07:00. Honestly, I'm impressed.

[an error occurred while processing this directive] >>> Back to post index <<<

Contact Email