Date: Sat, 13 Sep 2014 12:56:54 +0200 From: Polytropon <freebsd@edvax.de> To: Dave Babb <dcbdbis@comcast.net> Cc: freebsd-questions@freebsd.org Subject: Re: Request Validation of my Experience in buildworld Message-ID: <20140913125654.0ee1b126.freebsd@edvax.de> In-Reply-To: <54135498.8080202@comcast.net> References: <54135498.8080202@comcast.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 12 Sep 2014 14:16:24 -0600, Dave Babb wrote: > I think the idea of keeping the OS (buildworld) separate from ports > (portmaster -afdB) is nothing short of brilliance. I made several > mistakes in the process...as I am junior with FreeBSD...and at no time > was my OS taken down. I was able to gracefully recover and continue > until all was rebuilt. There is even another idea behind it: The "world", as well as the "kernel" are source trees where you can build everything or just a part, for example a modified version of fsck (and then install it) or a specific kernel module (and load it). Doing such things does not interfere with or create dependencies in the realm of application programs. And as you said, messing up your installed applications from ports does _nothing_ to the operating system. > After a buildworld, my system's memory footprint dropped by a good > 1/3...It sped up significantly. The I performed a "portmaster -afdB", > and rebuilt all my ports and their dependencies the long way. It's possible to add more compiling optimization in relation to what actual hardware your system runs on. Of course _today_ this is not really needed with our plentycore processors with gigs of RAM and endless hard disk, but if you have a limited hardware platform and _still_ want to have a responsive and stable OS, this is the way to go. Similarly, such optimization can be applied to ports. > Another dramatic drop in resources and another increase in performance. That is normal. Whenever I update my FreeBSD OS _on the same hardware_, the result is a _faster_ OS. This is not common in other areas of the PC world; in fact, the opposite seems to be true: you need to increase the hardware resources and hope to keep the same speed, but in most cases, you get a slower system with every update. > This experience was had across a desktop, and repeated on a ASUS laptop. > > Is this type of improvement typical? >From my point of view: Yes, it's the typical behaviour. That's why you don't need to fear a new major OS version, for example the upcoming 11.0. You don't need to buy a new computer. Instead, what you have will run faster and better with the new OS. The followint is true: Some years ago, I had a small server, Pentium 1, 150 MHz, with 128 MB RAM. It ran FreeBSD 4. Even though an update wasn't needed (because it kept working as intended), I decided to try the new (at this time) version 8 on a separate disk. Then I installed the server software that was needed and adjusted the settings from the previous install. I let the new system come up and... everything worked better! The server is still working today. I'm even planning to upgrade it to FreeBSD 10 at the end of the year, just for fun, and to see if the above theorie keeps being true. By the way, the machine in question was my _first_ FreeBSD desktop, running FreeBSD 5 at my home (before it became the FreeBSD 4 server, version 4 because the disk was simply taken from another machine that failed). That little box was able to run a WindowMaker desktop, play MP3s via xmms, also playing video with mplayer (built with optimized flags), run a FTP download, burn a CD (yes, no DVDs at that time) and build the kernel - while still being usable and responsive in the Opera web browser. Re-read the specs mentioned above! And people today complain about skipping audio when they move a window across the desktop... :-) -- Polytropon Magdeburg, Germany Happy FreeBSD user since 4.0 Andra moi ennepe, Mousa, ...
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140913125654.0ee1b126.freebsd>