I mentioned this before in the Lazy Reading from last Sunday, but it’s worth a second look: Apple’s new Fusion Drive product appears to be very much like DragonFly’s swapcache. DragonFly doesn’t have exclusive right to the idea of caching on a faster disk, clearly, so I’m not complaining that it’s “ours”. It’s frustrating to see product announcement/press releases stumbling all over this like it’s a new thing.
Then again, having new ideas about technology ideas and making sure they spread is one of the points of the BSD license, so perhaps there’s no good reason to complain at all.
(Before anyone reads too much into this: No, I don’t know of any direct relationship between swapcache and Fusion Drive; they may have no common background other than structure.)
A thread on pkgsrc-users@ reminds me: adding a specific line for bin-install will save time when rebuilding packages; pkgsrc will use existing binary packages instead of rebuilding from source when possible, when this is set. At least, I’m pretty sure that’s what it does.
Sandip Jadhav asked if anyone was working on an I/O scheduler. Chris Turner replied with a “no”, but also with a list of places to look for details on writing one, which I’m linking here for posterity.
John Marino is working on a very good idea: bringing bmake into DragonFly as a replacement for the current ‘make’. bmake is going through more active development and apparently also in use/will be used? on FreeBSD, so syncing up with the same make flavor as FreeBSD and NetBSD will help everyone. It’ll also remove the problem where you ‘make’ everything in DragonFly, except pkgsrc packages which you ‘bmake’. It’s not changed over yet.
(What does OpenBSD use for make?)
A conversation about compilers in the DragonFly base system led peeter (must) to describe his group’s use of OpenMPI on DragonFly for physics calculations. Apparently he’s had a significant performance improvement on DragonFly.
Along similar lines, John Marino helped out by bringing in libssp and libgomp for gcc 4.7 for use with OpenMP. (This is in DragonFly 3.3, not 3.2).
John Marino did a bulk build of pkgsrc using gcc 4.7.2, and posted the results. The result? About 1% of packages that built with gcc 4.4 did not build with 4.7.2. Whether that’s a problem with gcc or a problem with how each of those software packages were created by the original authors, I don’t know.
If you’re on master (or the 3.2 branch) of DragonFly, you’ll want to do a full buildworld/kernel. There’s been a pmap bug that looks like it’s fixed, but the changes have some repercussions. It may be possible to just recompile kgdb, but I like to overcompensate.
Google is hosting a ‘Doc Camp’, where people get together and write documentation for open source projects. There’s a page that talks about it. Last year’s inaugural event was apparently quite successful. I haven’t been to it, but I think a day just for documentation is a good idea.
I’m planning for DragonFly 3.2 to come with pkgsrc-2012Q3, the most recent release. I’m building binary packages to match, and the build should complete by the time we release on the 22nd…
Notice I said “should” – sometimes the universe conspires against bulk builds.
I branched 3.2 tonight. That means 2 weeks until release, so sharpen your bug-poking sticks!
(I’m very tired and unable to think of good analogies, sorry.)
Cause it could be added. The new algorithm could replace SHA-2, in use now in DragonFly. SHA-2 has not been ‘broken’ yet, so it’s not an emergency… yet.
I recreated the by-month thread and date listing from the old mailing lists, but for Mailman. It’s at lists.dragonflybsd.org.
Since the most recent branch of pkgsrc has been released, perl5 in pkgsrc has been updated to 5.16.1, and (ancient) python 2.5 has been removed.
Debian squished with DragonFly, sorta like Debian/kFreeBSD? Don’t know if it will work, but what the heck.
As I typed elsewhere, my general plan is to branch DragonFly 3.2 on the 8th, and release on the 22nd. That should give the recent scheduler and gcc work a chance to settle, and perhaps get a new version of USB support in too. It will probably be using pkgsrc-2012Q3, also, though we may not have binary i386 packages. 3.2 is shaping up to be a much more significant release than I expected.
The machine that runs www.dragonflybsd.org and bugs.dragonflybsd.org is currently down. While it gets figured out, Alex Hornung has a static copy of the dragonflybsd.org main website available.
Sepherosa Ziehau has some suggestions for anyone looking for some kernel hacking. They’re mostly based around busdma(9).
dragonflybsd.org appears to be down right now so I’m linking to the MARC kernel@ post.
There’s a post on the mailing list email@example.com of currently broken packages for the next quarterly release. It’s not a lot of stuff, but if something you need is on there, don’t worry too much. If you follow the thread through its replies, there’s a lot of fixing going on.
MARC, which stands for Mailing list ARChives, has a lot of mailing lists. It now includes the DragonFly users@ list, along with the others. (It’s not linked in *BSD on the main MARC page yet, but it should be soon.) It’s worth digging through the massive, massive wall of text on that page to find a mailing list you didn’t know existed.
This latest commit for the new scheduler means that on your next update, you will want to build a new kernel, and probably a new world too. This only applies if you’re running DragonFly 3.1, of course.
I got the old mailing list archives converted to Mailman. As I wrote in a post to users@, please let me know about problems. There’s some garbled messages from the old archive that were placed into the 2012-Sept. section for each message; I’ll be cleaning those up manually.
The old mailing list software for @dragonflybsd.org mailing lists, bestserv, apparently allowed people not subscribed to a list to post to it, after answering a confirmation message for each message posted.
The closest way to duplicate that for Mailman is to sign up for the list you want, and then turn off mail delivery for your email address in the config page for that mailing list. This won’t affect a lot of people, since most people want list output in their mailbox, but there’s at least a few I’ve fixed that way.
A discussion of why root automatically lists dotfiles with ls and all other users do not led to a long thread that includes some UNIX history. There’s some useful and some not-so-useful parts in the thread, but it did indirectly produce a way to reverse the listing effect itself.
Francois Tigeot benchmarked the recent Postgres 9.3 release. Postgres apparently switched to using mmap instead of SYSV shared memory, and Francois has done this to show the performance differences. (view the PDF in his post.) Of course, work has continued since this was posted, so there should be new numbers soon, and new changes I’ll document in a future post.
I haven’t found a reference to the exact decision Postgres made on how to handle memory; please post a link in comments if you know a good source.
See the note on pkgsrc-users@. The next quarterly release, pkgsrc-2012Q3, should be fully baked by the end of the month, if all goes well.
As seen in this pkgsrc-users@ post from Thomas Klausner, the freeze for pkgsrc-2012Q3 starts on Sunday and continues for (probably) two weeks before the release.
NYCBUG, the NY BSD user’s group, has an RSS feed for their speaker events, found via Dru Lavigne’s always useful BSD Events twitter. The next event at the start of October is a talk about SMPng in FreeBSD. Given that it was the project that in part led to the creation of DragonFly, I’d like to hear about it. (and even better, have someone more qualified than I compare and contrast that approach with what’s in DragonFly.)
If you look at new.pkgsrc.org, you will see what may become a new site. This is apparently a test, so don’t react as if this was the actual site.
If you ever wanted to read an extensive discussion about the scheduler, today’s your day. Mihai Carabas, who posted the details of a long discussion he had with Matthew Dillon about how the scheduler works. You may recall Mihai’s name from the very successful GSoC scheduler project that recently finished.
(look, a link to the new Mailman archive!)
All the mailing lists at @dragonflybsd.org have been converted over to Mailman. The old archives are still functioning, and will continue to update until I can find enough old material to retroactively complete the Mailman archives.
If you’re on any of the dragonflybsd.org mailing lists, I’m converting them over from bestserv to Mailman. I’ve done bugs@, commits@, hammer@, and test@ so far, and I’ll move the old archives over to the same format as soon as I find an actual mbox file with the old messages in it. The remaining lists should be tomorrow.
(If you got a note tonight from a list you were sure you were unsubscribed from, that was my fault; sorry! I didn’t understand the format of the bestserv user lists.)
DragonFly user varialus has created a page on the DragonFly website (it’s a wiki, after all) with all the notes taken from trying installation, etc. There’s far more notes than I expected there, so it’s worth a read.
I’ve uploaded DragonFly 3.0.3 disk images, both ISO and IMG. They should start appearing on a mirror site near you in the next 24 hours. This took a while after the tagging, I know, but I wanted to make sure every one of them booted. I didn’t on a previous release, and regretted it.
Francois Tigeot benchmarked several different operating systems using Postgres 9.2b3, including DragonFly, and published the results. I have a local copy of the PDF since the attachment didn’t really survive the archiving. Follow the thread for discussion. The Linux results look abnormally high, so it is possible that something different is happening on that platform…
DragonFly had a successful Google Summer of Code even this year. It marks our 5th time participating, 7th if you count Google Code-In events.
Mihai Carabas worked on adding SMT/HT awareness to the DragonFly scheduler. This project was very successful. The original goal was just to take advantage of threading with the scheduler, but the benchmarks published by Mihai show in general a 5% speedup from these scheduler changes. His work has already been committed.
Vishesh Yadav implemented an inotify interface in DragonFly. inotify is an originally Linux-based system for monitoring files and directories for changes. A specific use for this is an inotify-aware locate program, so that a list of file locations can be kept ‘live’. His code for the inotify interface should be committed to DragonFly very soon.
(This was written in part for Google to use on their Open Source Blog.)
I’m working on building new images, but: DragonFly 3.0.3 has been tagged. If you’re running 3.0, you can update and get some of the recent bug fixes.
3.2 is the next major release of DragonFly, which will be relatively soon by the every-6-months release schedule. John Marino’s put together another catch-all bug report for that release.
Sascha Wildner’s been working on his own DragonFly live images, in DVD or USB form. It uses XFCE along with a number of other packages listed in his post. They are .xz compressed, so they are nice and small for download, but make sure you have something that knows that format.
Mihai Carabas has posted his weekly results, showing a 5% improvement in pgbench resultswhen using his scheduler. Vishesh Yadav is working on IN_MOVED_TO/IN_MOVED_FROM flags (part of inotify, I assume). Ivan Sichmann Freitas I haven’t heard from yet. (Ivan, where are you?)
Juraj Sipos wrote me to describe MaheshaDragonFlyBSD, a live DragonFly image that has additional software preinstalled, and can easily be set to understand Sanskrit. It’s available in DVD and USB versions.
Here’s the regular status updates for Mihai Carabas (scheduler) and Vishesh Yadav (inotify).
I don’t have the update from Ivan Sichmann Freitas yet. Here’s Ivan Sichmann Freitas.
If you want to put something towards DragonFly, and you don’t have time or hardware, cash is now an option. (It’s not tax-deductible.)
The usual weekly updates from Mihai Carabas, Vishesh Yadav, and Ivan Sichmann Freitas. Mihai has some interesting bugs found this past week by running his code on Matt Dillon’s 48-core system.
The release announcement for pkgsrc-2012Q2 is out. New in this quarterly release: statistics about clang and pkgsrc. A surprisingly large number of packages build just fine with clang instead of gcc.
Attention students and mentors: the Summer of Code midterms open up on July 9th. This means students fill out an evaluation, and mentors also fill out an evaluation. Don’t forget, because completed evals from mentor and student both are necessary for a project to continue being funded.
More benchmarks, in this case a comparison of Debian GNU/kFreeBSD and DragonFly. I’m not even sure how to derive meaning from it.
Mayuresh Kathe asked about donations to DragonFly. I answered, but the part to remember is this: donate your time.
It’s almost an all-Vim week.
Your unrelated link of the week: Muppet Bohemian Rhapsody. Related: What kind of Muppet are you?
If you have an Intel processor with multiple cores and hyperthreading support, you can compile a new kernel and try out Mihia Carabas’s GSoC work already; he’s created a test using the OpenSSL test case to time scheduling performance vs. number of threads.
I know I already posted that this was on the way, but this time, the quarterly pkgsrc freeze is starting with a detailed announcement. 2 weeks until the next release, if everything goes well.
If you’re using some PHP application that requires the old behavior of PHP 5.2, you will need to specify that version of PHP – pkgsrc is moving to version
5.4 5.3 as default, with version 5.4 available. (thanks, Takahiro Kambe for the update.)
The freeze for pkgsrc-2012Q2 starts on the 16th of June, as recently announced. Freezes are usually 2 weeks, so that means 2012Q2 should be tagged at the end of June.
Pkgsrc already runs on a large number of different platforms, but that’s not what I’m talking about. In this case, Joyent, which uses pkgsrc internally, has a suggested change that makes binaries usable on both 32 and 64 bit systems. I don’t know if this will go into pkgsrc proper, but it’s interesting to see.
Apparently a lot of modular-xorg packages in pkgsrc received updates. I think I found some of the changes, but probably not all, so I don’t have a good way to sum up the actual effect.
Update: see the end of this cvsweb pkgsrc CHANGES-2012 page for all the changed parts.
There’s a number of packages out there that assume you are using the GNU versions of ls, wc, and so on. However, you aren’t when using a BSD system. Pkgsrc has historically dealt with this when GNU tools are needed for a package by prefixing them with a ‘g’. ‘ls’ becomes ‘gls’, and so on. Aleksey Cheusov proposed a fix to keep these utilities under their original names, which I think will go into the next quarterly pkgsrc release.
Pkgsrc packages that have source files that can’t be redistributed, and go missing for the length of an entire quarterly release, will get removed. They are effectively broken at that point anyway.
That policy is now formally in place; I don’t think there was a clear prescription before.
Venkatesh Srinivas, currently on his colossal bike ride, introduced a different way of creating a tmpfs. This was test code, and Johannes Hofmann benchmarked it (see same page). It’s interesting cause there are numbers, and nice to see one person jumping in to test someone else’s results/idea.
I think I’ve mentioned building DragonFly with clang before, but not pkgsrc. There’s two variables to set, plus some special handling for libf2c. Thomas Klausner has details. This is not tested on DragonFly.
There were some benchmarks of DragonFly 3.0 some time ago on Phoronix. (You may recall it being mentioned here previously.) The disk numbers always seemed weird to me, so I repeated that part of the test, and here’s my writeup.
Takahiro Kambe is bringing PHP 5.4 into pkgsrc, probably as lang/php54. Follow the whole thread for a discussion of version numbering. As a side effect of this, PHP 5.2 will leave pkgsrc by the next quarterly pkgsrc release. If you’re using that older flavor, you’ll want to upgrade.
TUI mode is available now for kgdb on DragonFly, thanks to John Marino. It’s apparently a Text User Interface for debugging core files. I haven’t used it, so I’m relying on the testimony of others.
Michael Lucas has a writeup on how he debugged his RANCID setup. I link to it for the technical details, and also because if you have to manage more than a few switches or other network devices, RANCID is very useful.
Here’s a post by yours truly, on how to move to pkgsrc-2012Q1 though building from source. This is for anyone sick of waiting for me to finish the binary build of pkgsrc.
Matthew Dillon posted a followup on that fix for clustering I noted yesterday. It describes the exact problems better than I could, though the result is the same: you should update if you’re running bleeding-edge DragonFly.
A fix for cluster_write() issues reported by multiple people is now available, so if you’re running a version of DragonFly newer than 3.0.2, you’ll want to update.
Peter Avalos has updated OpenSSL, though this version is apparently a bugfix, not a security fix. Still need it anyway, since it disabled TLS 1.1 in an unexpected way. See the OpenSSL changelog entry at “[26 Apr 2012]” for details.
Each of the 4 DragonFly participants for Summer of Code have posted an introductory email and details of their projects. Here’s direct links to their posts for your reading convenience:
(Yes, same format as my last post, but now the links are to their posts, not the sparse Google info pages.)
There’s a few pkgsrc packages that might be going the way of the dodo, soon. There’s a few more that need love, so speak up if you use them. Maybe you can be the Somebody™ that fixes them?
Welcome our newest committer: Markus Pfeiffer. He’s ‘profmakx’ on EFNet #dragonfly, and has been working on a port of FreeBSD’s USB infrastructure – which I am looking forward to, tremendously.
Mosh, mentioned on this Digest a few weeks back, is now installed on leaf.dragonflybsd.org. If you’re doing any development work there but dealing with a relatively high latency, this should help. (Thanks Venkatesh Srinivas.)
I’m still working on building them. I kept getting panics, which seem to be fixed by this commit, so I should have something soon. Sorry!
Michael Lucas’s worthwhile book, SSH Mastery, is currently having one of those sudden price cuts on Amazon – for the paperback version, about 25%. Now it a good time to nab it before the price bounces back up.
DragonFly now has its own ntp.org zone. What’s this mean? Nothing material, but it’s nice to do.
That DragonFly review is now available in all six parts. (I included the preamble there.) I still haven’t made it through the whole thing.
Julian Fagir has put together a graphical – meaning it works under curses in a terminal, or under X - interface to pkgin, the binary package manager. Can someone try it and describe how well it works?
There’s several packages that will be removed from pkgsrc after the 2012Q2 branch, since they haven’t worked in a long time. Also, Python 2.4 has been removed from pkgsrc-current and 2.5 will go the same way before the end of the year.
The next quarterly release of pkgsrc, pkgsrc-2012Q1, has been branched. I’ll start building binary packages momentarily.
The branch should show up in DragonFly git later today. Once available, you can change any references to ‘pkgsrc-2011Q4′ in /usr/Makefile to ‘pkgsrc-2012Q1′, and then to switch to it:
- cd /usr/pkgsrc
- git branch pkgsrc-2012Q1 origin/pkgsrc-2012Q1
- git checkout pkgsrc-2012Q1
- git pull
At that point, you can start building and installing newer applications. For more details on that, check the pkgsrc guide on the DragonFly website.
Note that you don’t have to do that; you can stick with the 2011Q4 (or earlier) packages you have installed now, if you don’t want to deal with software changes right now, or if you want to wait for the binary packages to become available. Upgrades/security fixes only happen for the latest quarterly release, though.
Note: don’t assume I tested this before advising you to do it, or anything like that. I mean, come on.
Steven Rosenberg is writing the longest DragonFly review ever. Here’s parts one, two, and three. There’s 3 more parts to come, 1 per day, so check back for the end of the story.
There’s been some discussion of packages that have been broken for a long time in pkgsrc, over on the firstname.lastname@example.org mailing list. It’s interesting to see just what breaks these packages, though it still seems up in the air whether any will be removed or not. (Follow the thread if you have time.) I don’t think the discussion has ended yet.