My 10 Minute Experience With PC-BSD 10.0

[sulman]

PCBSD needs more then 2GB of RAM and 2GHz of CPU to run. And that's the absolute minimum. When they made ZFS the default file system, PCBSD now needs way more that and this is all just to run fluxbox.

The reason why PCBSD is so slow is because it is literally FreeBSD running KDE and other modern desktop environments. The FreeBSD wasn't designed to take that sort of load and it never will. Also, Systemd and Upstart out performs the BSD init by light years and it's kernel can only search and initiate drivers serially (unlike Linux which initiate all in parallel).

I recommend not switching to PCBSD because it is so buggy, bloated and it has a lot of services turned on by default which coupled with the lack of ASLR, make it a serious security liability.

I especially don't recommend switching to PCBSD 10.0 because of the tendency for the X.0 versions to be very very buggy.

This isn't correct. You can run FreeBSD with a heavy desktop environment and it's pretty much identical in experience to anything else.

It's not going to be as fast on start as SystemD, but not much is nowadays; just about any Linux spin not using Systemd or Upstart is about the same. If a five second boot matters to you then obviously Debian/Gentoo with Openrc/FreeBSD isn't going to cut it.

[JX8p]

PCBSD needs more then 2GB of RAM and 2GHz of CPU to run. And that's the absolute minimum. When they made ZFS the default file system, PCBSD now needs way more that and this is all just to run fluxbox.

The reason why PCBSD is so slow is because it is literally FreeBSD running KDE and other modern desktop environments. The FreeBSD wasn't designed to take that sort of load and it never will. Also, Systemd and Upstart out performs the BSD init by light years and it's kernel can only search and initiate drivers serially (unlike Linux which initiate all in parallel).

I recommend not switching to PCBSD because it is so buggy, bloated and it has a lot of services turned on by default which coupled with the lack of ASLR, make it a serious security liability.

I especially don't recommend switching to PCBSD 10.0 because of the tendency for the X.0 versions to be very very buggy.

The first point is unsourced.

The second point is not based in any fact. No operating system was designed with a view to running modern graphics environments; most operating systems of major use today were built, at least originally, before the mid 90s. As regards Systemd and Upstart, Upstart is soon to come to GNU/kFreeBSD at least. Launchd's port to FreeBSD is also getting updated. As for BSD init's slowness, you really are laying it on a bit thick. It is barely slower than systemd or Upstart.

Third point is totally subjective. Most desktop operating systems have lots of services turned on by default. If you dislike this then run something that leaves little on by default: FreeBSD vanilla, Arch, Gentoo, and their likes. ASLR is an insignificant feature which is useful for protection against rootkits, but most of those already work around it. That is my opinion. But if you don't like my view, please find Oliver Pinter's ASLR patch for the FreeBSD kernel. As for 'serious security liability', are remote exploits a big issue for desktop users? Not really. Most of them are behind NAT which puts an end to that before it reaches their PC. Server is another matter, but PC-BSD is for desktops.

Fourth point has a little truth in it -- X.0 versions of FreeBSD and its derivatives can be buggier. However - and this may not be scientific, but it's a good measure anyway - the general consensus in #FreeBSD on Freenode is that this is a very stable release. In any case, that's no reason to advocate against switching to PC-BSD, and 'very very buggy' is an overstatement, even for buggier point-zero releases from the past.

[Guest]

3. I do not know why FreeBSD developers haven't adopted ASLR or an ASLR-type of security I'm not one of them to explain the reasons why but I'm sure there was much talk about it. ASLR first appeared in OpenBSD so I'm sure it would not constitute as big an effort to implement it as porting from somewhere else, Linux for example.

To be perfectly correct it was ported to OpenBSD two years after PaX made it for Linux. ASLR is WiP for FreeBSD:

http://0xfeedface.org/blog/lattera/2013-11-16/freebsd-aslr-progress

It's taking so long, because lack of resources and different priorities than OpenBSD in my opinion.

[Pallidus]

first world problems:


"I only got haswell laptops to test on"

motherfucker I will ship you a couple of old ass t60's and you can give me that haswell