FreeBSD 7.1 Released 324
Sol-Invictus writes "The FreeBSD Release Engineering Team is pleased to announce the availability of FreeBSD 7.1-RELEASE. This is the second release from the 7-STABLE branch which improves on the functionality of FreeBSD 7.0 and introduces some new features. Some of the highlights:
The ULE scheduler is now the default in GENERIC kernels for amd64 and i386 architectures. The ULE scheduler significantly improves performance on multicore systems for many workloads.
Support for using DTrace inside the kernel has been imported from OpenSolaris. DTrace is a comprehensive dynamic tracing framework.
A new and much-improved NFS Lock Manager (NLM) client.
Boot loader changes allow, among other things, booting from USB devices and booting from GPT-labeled devices.
KDE updated to 3.5.10, GNOME updated to 2.22.3.
DVD-sized media for the amd64 and i386 architectures."
Re:Netcraft confirms it: (Score:1, Informative)
tag it : reportsofmydemisehavebeengreatlyexaggerated
*Finally* DVD media (Score:5, Informative)
This is one of the better parts of this release. The lack of speed/clue on putting out both CD sized and DVD iso images has been highly frustrating, telling the users to basically "roll-their-own". I've already upgraded a few systems and things appear to be going well.
Description of some changes, links to benchmarks (Score:1, Informative)
Re:Java on FreeBSD (Score:4, Informative)
Re:Dont forget documentation (Score:5, Informative)
Re:Hmmm... (Score:3, Informative)
Netcraft is not reliable anymore. From the site:
Why do you not report uptimes for Linux 2.6 or FreeBSD 6 ?
We only report uptimes for systems where the operating system's timer runs at 100Hz or less. Because the TCP code only uses the low 32 bits of the timer, if the timer runs at say 1000Hz, the value wraps around every 49.7 days (whereas at 100Hz it wraps after 497 days). As there are large numbers of systems which have a higher uptime than this, it is not possible to report accurate uptimes for these systems.
The Linux kernel switched to a higher internal timer rate at kernel version 2.5.26. Linux 2.4 used a rate of 100Hz. Linux 2.6 used a timer at 1000Hz (some architectures were using 1000Hz before this), until the default was changed back to 250Hz in May 2006. (An explanation of the HZ setting in Linux.)
FreeBSD versions 4 and 5 used a 100Hz timer, but FreeBSD 6 has moved to a customisable timer with a default setting of 1000Hz.
So unfortunately this means that we cannot give reliable uptime figures for many Linux and FreeBSD servers.
Re:Hmmm... (Score:5, Informative)
FreeBSD fanboyz shouldn't go mouthing off about "half-assed" considering the way since 5.x it's crappy smp and threadlocking would seize up tighter than a great-grandma on a straight brick cheese diet with lock-mgr panics. Problem persisted in 7.0, who knows if 7.1 will finally put the issues to rest?
Are you talking about this SMP [slashdot.org]?
5.0 was released in January 2003, I think 6 years of passage should have allowed you enough grumping time that you can let it go now. I think you could also take a look in your wayback machine and remember that Linux was not exactly perfect at the time either. FreeBSD 5 did have its teething problems with all of the new technologies introduced, especially KSE and the ULE scheduler, but progress has continued to be made and your unsubstantiated claim otherwise is just the pathetic grumblings of a troglodyte.
Re:Hmmm (Score:3, Informative)
I'm sorry, those KDE and Gnome versions are some random software the FreeBSD people maintain that have nothing to do with the third party KDE and Gnome systems available on Linux?
How about SSH? You don't use the OpenSSH maintained by the OpenBSD guys then? This is some other version of SSH that is completely maintained by FreeBSD?
Stop with the FUD. Maintaining a local code base for applications doesn't change what they are. The Fedora people do the same thing, so does Ubuntu and many other distros. Slackware and Debian are pretty hands-on in their code maintenance as well.
You are correct (Score:5, Informative)
Everything in the ports tree is essentially random crap. The only thing FreeBSD does is wrap the source code with a (really nice) build system. Ports aren't "stable" the same way the core is. That said, a lot of the big-name stuff like apache has separate ports. For example apache-1.3, apache-2 and apache-2.2 have separate ports (I think there is a port that follows the trunk too).
The difference between the BSD's and Linux's are in scope. In FreeBSD there is a whole lot more junk that is maintained by a single group then in most linuxes.
But still, you are correct in the "real applications" are all ports.
Re:Hmmm... (Score:2, Informative)
A significant improvement on a crappy OS is still a crappy OS. No flame intended, but really... who uses FreeBSD anymore?
I do. On my servers as well as on my workstation and my laptop. It's a solid and stable operating system that has earned its place.
Apart from that, only some some [yahoo.com] smaller [nyi.net] companies [netcraft.com] use it.
No flame intended, but really...who critizes operating systems that are not sold by Microsoft on /. anymore?
Re:SMP + Stability = Win! (Score:5, Informative)
Hi!
I'm one of the Squid developers and I have some experience with FreeBSD :)
FreeBSD-6 and FreeBSD-7 both rock for Squid (and my squid-2 fork, cacheboy.)
FreeBSD-7 is pretty scarily scalable when it comes to web stuff. I'm working on threading cacheboy/squid-2 over the next few months enough to take advantage of the parallelism that the FreeBSD guys have introduced into -7 and -current. I've got some test code here for fully transparent web interception caching with FreeBSD-current, and some stuff to use FreeBSD's fantastic POSIX AIO support.
Its all lookup up, up, up from here. :)
Re:what about smb speed-ups? any?? (Score:3, Informative)
Re:Man pages are not a quality control technique! (Score:3, Informative)
What was that quote again?
Those who don't know UNIX are doomed to reinvent it -- badly
ZFS (Score:4, Informative)
FreeBSD is the only distribution, other than Solaris, to have ported and implemented the ZFS filesystem (and no, a FUSE port doesn't count).
I've been looking forward to build a file server for personal use, and I'm eager to try out ZFS, which really puts FreeBSD high on my small list of candidates for an operating system. I'm going for consumer-grade hardware, and I'll be experimenting with stuff like using CompactFlash cards to store the OS.
OpenSolaris was my initial choice due to its higher maturity on the ZFS implementation, but I feel it's too constraining. I tried searching around for information about installing the system on flash mediums, information about wear-levelling, filesystems for flash media, and their forums and mailing lists fall short on these topics. The OpenSolaris installer doesn't even allow one to customize the installation, forcing me to install X.org, Gnome, and a ton of other stuff. No thank you, I'd very much like my file server to be command-line only, and to be smaller that your 3.1 gigabyte minimum for an installation.
As soon as I feel that FreeBSD's implementation of ZFS is stable and feature-rich enough for my needs, I'll definitely be rolling a file server with it. And I don't care if Netcraft disagrees with my decision; I really do feel BSDs deserve more and more notoriety these days.
bits from the trenches (Score:1, Informative)
I've been deploying FreeBSD in the last 10+ years, mostly as chief architect/CTO in startups, and here are some of the reasons I still opt for it today - currently about a dozen servers in cages serving cumulatively 200 Mbit/sec, all non-http proprietary protocols and software (we have nothing web-related,) software-based redundancy, MySQL, etc:
1. Stability.
In the last 10 years I had kernel panicked once, due to a driver problem.
This needs some discipline and patience:
- We introduce a new RELEASE only after we have been running it for at least 6 months on our test systems, and then only when it becomes cost-effective over patching the previous one.
- Everything we run is compiled from the source, to the last executable, on the target architecture (we have several.)
- We create our own automated installation media, certified for our targets, which results in half-page instructions and few CDs that a retard can install in less than an hour - from bare metal to the running server.
- We never run any GUI on servers. All access is via SSH (which is actually the only open port apart from our proprietary protocol ports) and engineers use a wild range of development/interface machines. Some even use Linux.
2. Maintenance
Usually, for each release, there are few problems that we have to solve.
Most of the time it's patching network drivers (as hardware is always ahead of the release). Search for the problem, figuring out what to do, testing the patch, incorporating it into our installation process - all of that never took more than 2-3 days. At this point I am willing to bet that 7.1 will run on all our servers with less than a week of extra work (currently we are on 7.0. 7.1 will get phased in in August or later.)
Outside kernel, there are sometimes application bugs, generally nastier than kernel problems. MySQL had one with faulty locking on multiprocessor machines, which we patched in May last year (it took about a week), and which only recently found its way to MySQL release.
All taken into account, it's usually less than 2 weeks of work and about 6 months of test run for each new release.
3. Performance
We never ran into a performance issues, simply because we add a server when average CPU/io usage goes over 20%. Our job mix appears to uniformly load CPU, network interface and disks. A server costs $4-8K and it's a non-issue for the kind of business we run. So maybe we are just lucky here. The only real issue is bitmap rendering that we do in batch once and a while (can take a week on *fast* 4-core server), but we do this once every few months.
None of us, except one strange guy, uses FreeBSD as workstation machine. It sucks to force Linux GUIs onto FreeBSD (even the weirdo runs some simple screen splitter), so I have no clue about that aspect.
4. Elite status
They used to say, "Would you rather date crowd in PC or Mac section in Frys?". Well, you get the idea.