Matthew Dillon happened to write something about RAID vs. SATA drives that says, among other things, picking SCSI over SATA drives doesn’t make financial sense.
Matthew Dillon started to think about writing his own bug tracker, to which the genral response was “Keep up the cool DragonFly work!”. Many people are leaning towards Jira, though Hiten Pandya still wants to evaluate Bugzilla or cvstrac if Jira does not work out by the time of the next release.
Matthew Dillon, like many people, thinks hardware RAID is generally better. In fact, he’s using cards from 3ware.
Also, Simon ‘corecode’ Schubert looked at the bug tracking program jira, and liked what he saw. He liked it so much, he set up a test installation.
The BSDCertification website has a new survey up. Unlike the last one, it’s only 19 questions and very short. It’s available in a number of languages.
Jeremy Messenger suggested this comparison list for bug trackers; Hiten Pandya also wants to look at MySQL’s Eventum.
Brad Schonhorst kindly sent along this announcement about NYCBSDCon:
“There are only 2 days left to preregister for the first ever NYCBSDCon, to be held September 17th at Columbia University, in the Davis Auditorium.”
There’s more at the NYCBSDCon site. Among other notable speakers, our very own Jeffrey Hsu will be presenting “History, Goals, Objectives, and Structure of DragonFlyBSD”. I’d really like to go see this.
UnixReview.com has a review of the book “Shell Scripting Recipes: A Problem-Solution Approach“, some writeups of what some people think of computer certifications, and some date-related shell functions.
Joerg Sonnenberger noted there are some useful steps one can perform for transition to pkgsrc.
Author Neil Gaiman found a tomato that looks like it has a horn; add a matching second horn and it would look like Beastie.
Incidentally, Neil Gaiman has little to do with BSD, but he has written some excellent books and comics.
The systems mentioned in a previous post as being ordered for dragonflybsd.org will be “ASUS A8N-SLI Deluxe”, which means that they will probably be good choices for supported hardware.
Matthew Dillon’s ordering some new hardware for dragonflybsd.org. Among other improvements, there will be a new machine specifically for building world, kernel, and pkgsrc.
On users@, Andreas Hauser talked a bit about his experience with filesystems, and included some links. Among other things, he pointed at the possibility of NFSv4.
Matthew Dillon listed his two major UFS goals: changing filesystem size, and speedier reboots.
Tomaž Borštnar used ubench to test DragonFly and various FreeBSD systems, and wrote up the results.
Speaking of benchmarks, the fefe.de benchmarks done some time ago may be coming in for a second round of tests. Of course, that blog entry is in German, so I’m going on what Hubert Feyrer said.
If you’ve been running a DragonFly mirror, the BSD Certification Group would like to know your download stats. Several people have quoted healthy numbers. (post them to kernel@dragonflybsd.org, if you have them)
Matthew Dillon wrote a little more on why he thinks pkgsrc is the best direction for DragonFly.
He also followed up on a separate thread describing the current disk size limits for UFS1. Hiten Pandya added comments, too.
Simon ‘corecode’ Schubert proposed using cvstrac for bug tracking, and he has set up an example. Scott Ullrich has already been using cvstrac for pfsense. However, Hiten Pandya was intending to set up Bugzilla.
Simon ‘corecode’ Schubert now has release, development, and preview versions of DragonFly built regularly on his server. In addition, he has the most recent version source for each type available as a .tar.bz2 file, for those who don’t want to grab it through cvsup.
Jeremy C. Reed posted his results from building all possible pkgsrc packages on DragonFly. More than half were successful already, and more (since he was building as non-root) should be possible. You can read his full post for details, but here’s a nice little summary:
Build started: Mon Aug 29 19:45:33 2005 GMT
Build ended: Wed Aug 31 03:36:14 2005 GMT
Successfully packaged: 2902
Packages really broken: 1037
Pkgs broken due to them: 1270
Total broken: 2307
Not packaged: 314
Total: 2621