A running description of activity related to DragonFly BSD.
Sascha Wildner has added an option to the installer to create a UFS boot and Hammer volume as an install disk, in addition to the all-Hammer and all-UFS options already available. Programs expecting the booting kernel to be on UFS will be able to find it, but users still get the benefits of Hammer.
Updated: It replaces the all-Hammer option. Thanks for the correction, Sascha!
Matthew Dillon has a new version of Hammer, which speeds up listings from programs like ‘ls -la’ and ‘find’. This is only in 2.3.1.x code right now, so don’t force an upgrade via hammer version-upgrade if you’re still on DragonFly 2.2. His post includes some benchmarks.
On a side note: sili(4) tests look good.
As Jim Chapman found out, dump only works for UFS, and not for Hammer. Matthew Dillon outlined the different mirroring and snapshot methods that Hammer makes available.
I’ve heard of Dvorak keyboard layouts, but I didn’t know there’s another, called Colemak. Sascha Wildner has committed a patch from Geert Hendrickx which makes Colemak layouts available on DragonFly.
Matthew Dillon has changed the way USB flash drives are attached, to make sure they don’t interfere with AHCI-attached disks. This is temporary, and will be replaced by a dynamic /dev.
Siju George described his efforts to set up a continuous, automatic backup system using Hammer, with some interesting results. Matthew Dillon chimed in with some suggestions.
Matthew Dillon is looking for one more driver to build to complement the AHCI and Sili drivers. There’s several suggestions already.
Matthew Dillon’s made some small changes to Hammer; it should result in a small speedup when copying data.
It’s the weekend, so it’s a good time for a digression. This blog post from Matt Trout describes a lot of the code work he’s done for Perl, and what he thinks the best contribution is. The important part is the end of the post. He notes that for all the code he’s added, the best return has come from encouraging others to contribute. The net result has been a magnification of effort, as more people donate time.
The reason I’m posting this is to note that DragonFly, as a community, has been excellent so far at providing a low-drama environment for people to have ideas and contribute work. Keep this in mind; the best benefit to DragonFly isn’t lines of code, but people welcomed.
‘Haidut’ brings word of a 50-system DragonFly installation acting as web crawlers, with performance exceeding that of the Debian Linux systems they replaced. There’s more details about what’s being run, if you’re curious.
A number of people have noticed that Hammer’s pruning (which by default runs once at midnight) makes systems temporarily unresponsive. Matthew Dillon’s committed a fix for this, with warnings of more improvements to come.
I recently did a bulk build of pkgsrc on two similar machines; the only significant difference being extra CPU work being done on one system, and Hammer snapshots on the other. However, they’re diverging in speed over time, which is interesting but not yet conclusive. Read my post about it for more details.
A good benchmarking project would be testing Hammer with snapshots on and with snapshots off.
Taking from his AHCI work, Matthew Dillon’s working on a Silicon Image 3132 driver. An initial version is available now, though the usual caveats about a brand-new device driver apply.
Update: he’s really moving fast on this.
Hasso Tepper posted his notes on the pkgsrc-users@ mailing list about the different video modes for the Intel video driver. Version 2.7 works, but only if you use certain options.
The pkgsrc freeze is on. We should have release 2009Q2 in 2 weeks…
If you’re one of the few who has seen a ‘no local apic!’ error when booting, Sepherosa Ziehau’s recent commit may have a fix for that. He asks for testers, though he cautions to do it without APIC_IO in your kernel config.
Matthew Dillon is relentlessly adding to his AHCI work, with a new status report summing up the speed and stability improvements. The driver will probably end up in the next DragonFly release.
Matthew Dillon has initial support in for port multipliers, along with other AHCI work. It’s not ready for production yet, and he lists the various issues going on, including a need for a different way to mount disks – AHCI changes devicenames from ‘ad’ to ‘da’, which can be a hassle.
Update: hot-swap support, too.
Update update: parallel scans for speed.
The freeze for pkgsrc’s 2009Q2 release starts this Sunday, the 14th. The 2009Q2 release should follow two weeks afterwards, which will be very close to the time of the next planned DragonFly release. (2.4, in case you weren’t counting.)
I’ve just finished a new build of the 2009Q1 packages for DragonFly 2.2, and it’s available on http://avalon.dragonflybsd.org/packages – setting BINPKG_SITES or using pkg_chk can get you these latest versions.
I plan to have a 2009Q2 package set for DragonFly 2.4 as soon as possible after release.
Sascha Wildner has posted a patch that makes it very easy to switch out the compiler used to build DragonFly. This builds on earlier work from Alex Hornung.
This should make it into the base system. Everyone’s looking at compilers that aren’t gcc these days, it seems.