In honor of my trip to Hawaii, here’s a screenshot with GNOME’s Fraser Island background. I just switched to it, and it got me in the mood for a screenie.

I recently started using Graveman to burn CDs. It has a simple, task-oriented interface that made it easy for me to get my downloaded ISO onto a CD.

Try it out.

We had a discussion tonight on the uses of our newly released installer and how it could help increase enterprise use of Gentoo. One of the things that came up was using it as a plugin to some sort of Web-based management framework along the lines of RHN. Immediately we thought no reason existed for such a tool to be distribution-specific; as needed, it could have distro-specific plugins as gnome-system-tools does.

Nobody on #gentoo-installer knew of an open-source project designed to install and administer entire networks of machines along with monitoring and auditing of the network itself, so I went looking for one. The ideal system would have a plugin-based architecture, so the administrator could decide how much power the Web interface had. This would reduce the potential damage from screwups or security holes in the interface. Plugins would be available for installation of systems, installation and update of either specific packages or the entire system, backup/recovery, network monitoring and network auditing.

The goal of such a project is to have a single tool for creating and managing the network holistically, including the individual computers within the network as well as the network itself. This would reduce the work necessary for the network/systems administrator by centralizing and simplifying network maintenance.

What did I find?
m23: A Debian-specific tool for installation and administration of “hundreds of clients.”
LRS — Linbox Rescue Server: Centralizes images, backups, etc. Can restore with PXE, multicast TFTP, or using a bootable medium.
LCM — Linux Cluster Manager: A Beowulf cluster installation/management tool that can run commands, show stats, and do system imaging.

That’s about it. Anyone seen more?

Update: I’m talking primarily about management and configuration tools, not installation tools. Don’t tell me about kickstart etc, I know about it.

As I was searching freshmeat today for a webapp that could manage networks of computers effectively (something like RHN, with an option to plug in network management tools for monitoring, auditing, etc.), I came across a tool called remerge.

If you know anything about Gentoo, you’ll be able to guess that it supports us. But it also appears to support Fedora, SUSE and NLD. Oddly, this app wasn’t in portage, despite being at version 2.1.1 (not exactly unstable-sounding), and there also wasn’t a bug filed for it.

The Fedora claim seems a tad shaky though, since what it really supports is Red Carpet and nobody really uses RC on Fedora.

I’d be interested to hear whether people have tried out remerge, and their experiences with it.

Something like this would be quite useful in gnome-system-tools, instead of each distribution producing tools like synaptic, whatever Fedora’s GUI is, the package manager within Yast2, etc.

Same score as dang. Wonder what that general knowledge question is that we both got wrong, assuming Genius is the top score in every category.

The author of a recent O’Reilly book, The Art of Project Management, just posted on article titled “How to Decide What Bugs to Fix When, Part 1.”

Much of this article is quite easy to apply to Bugzilla, which suggests to me that either all good bug databases have similar features or the author likes the Bugzie.

For example, he says we should split bugs into multiple priority and severity levels, combining both when deciding whether to fix something. He defines priority like this, with a 4-level scheme:

• Priority 1: Must fix
• Priority 2: Might fix
• Priority 3: Probably won’t fix
• Priority 4: Won’t fix until hell freezes over

And then we can use severity to define how major a bug is when it hits somebody (combination of my definitions and Bugzilla’s), regardless of how often it hits people:

• Blocker: Breaks their systems, loses their data, requires reinstallation
• Critical: Crashes, loss of data, severe memory leak
• Major: Major loss of function (e.g., an e-mail client where you can’t send e-mail)
• Minor: Minor loss of function, or other problem where easy workaround is present
• Trivial: Cosmetic problem like misspelled words or misaligned text

To decide which bugs to fix, we first sort by priority to get a list of bugs we even care about. Then, we sort by severity to know where to start.

For anybody with an overwhelming bug list, you might like to start using Bugzilla’s priority field. I know I plan to.

Here’s my post to the gentoo-dev list, for anyone who doesn’t read it and does read this.

It can’t recognize duplicate posts. Every time I blog, I see my post five separate times from four separate Planets and my own blog, never marked as read.

Using fun scripts like this, I’m checking and adding packages:
for i in *; do if [[ "$i" != "xcursorgen" ]] && [[ "$i" != "xkbcomp" ]] && [[ "$i" != "CVS" ]]; then cvs add $i ${i}/files; find $i -type f | grep -v CVS | xargs cvs add; pushd ${i}; echangelog "Initial commit for modular X."; cvs add ChangeLog; repoman scan; popd; fi; done

Just switched over to Liferea from Blam. It has folders, so I can categorize my feeds by their priorities — on busy days, I can just read top-priority feeds, etc.