i prefer to use touchpad i switched back to linux now. if some day i wanted to go back to FreeBSD i will open a forum topic.

i prefer to use touchpad i switched back to linux now. if some day i wanted to go back to FreeBSD i will open a forum topic.

That’s sad... but understandable

Hi, I'm new to FreeBSD, though, tired it long back. I've a new installation with Gnome-3 and display works just OK with drm-legacy-kmod. My doubt is, if 2nd gen IGP (Sandybridge graphics) is supported in any new drm drivers? I tried drm-kmod and no display!

Hi, I'm new to FreeBSD, though, tired it long back. I've a new installation with Gnome-3 and display works just OK with drm-legacy-kmod. My doubt is, if 2nd gen IGP (Sandybridge graphics) is supported in any new drm drivers? I tried drm-kmod and no display!

https://wiki.freebsd.org/Graphics

Welcome 🙃

Welcome 🙃

Thank you.

I read it. And, drm-legacy-kmod seems to be the intel driver available for old IGPs.

Hi. I've installed FreeBSD on a headless raspberry, and I'd like to install a bunch of programs from pkg, minimal way. For example, I'd like to install mutt, but I don't care of libgmp. Is it possible to install a minimal mutt?

Do I need to compile my own via ports, for that to happen?

It's preferable (and easy) to use ports : - it will adapt to your environment - you can access to installation options

It's preferable (and easy) to use ports : - it will adapt to your environment - you can access to installation options

I like the idea of having a very custom set-up... but this would also mean that I have to recompile the software on my own, is that correct?

In the manpage I don't see an explicit mention to different flavours. Do you mean that the repositories have all the combination of build-time configurations for all packages? It sounds weird: it is probably a huge number of packages with weird names! :)

I like the idea of having a very custom set-up... but this would also mean that I have to recompile the software on my own, is that correct?

With ports you always compile the software by your own. What exactly are you planning with a "custom set-up"?

With ports you always compile the software by your own. What exactly are you planning with a "custom set-up"?

The idea is to keep everything minimal (e.g. no need for a png support, or mouse support? - do not add it). It is desirable not to compile everything, as the raspberry is not very powerful, so I like to use pkg, but then I lose the minialism. It might be ok to compile somewere else, but then I should have some cross-compilation towards ARM, as it should eventually run on the raspberry.

I see! Thanks for the hint, I'll check!

Helloo

Hi there!

Hello!

Hello

İ have question. İ use unbound for dns web filtering. Unbound restart when i write rule then internet connection is gone for 20 - 60 sec. i dont want that how can i change unbound cache without restart? Or how can i use internet when unbound restarting? Someone help me?

İ have question. İ use unbound for dns web filtering. Unbound restart when i write rule then internet connection is gone for 20 - 60 sec. i dont want that how can i change unbound cache without restart? Or how can i use internet when unbound restarting? Someone help me?

Use two unbound and don't restart simultaneously 😉

Use two unbound and don't restart simultaneously 😉

Thank you, i thought that but it use so many ram. İm looking for method or something in unbound

Nice thank you

It's preferable (and easy) to use ports : - it will adapt to your environment - you can access to installation options

Hi Phil, pls let me ask one question of understanding: is it still not recommended to 'mix' pkg/ports on the same computer? If so, do you know a technical answer for this? Thx.

I read somewhere in bsd-related forums not to mix the package system with the ports system...

Hi I also read that and to be honest with you I really don't know..

I had a server where I only compile from ports and works perfectly

And on my personal computer I had mix between pkg and ports it works ... The last problem that I get was a mismatch version of qt .. but I fixed it

In the end pkg will import you the compiled port entity

And on my personal computer I had mix between pkg and ports it works ... The last problem that I get was a mismatch version of qt .. but I fixed it

I mixed ports and pkg on my server That lead to problems once I tried to update ports

I think if you compile from ports for a desktop with all libraries and some program aren't lightweight like Firefox .. or some of them ask you to install lib that are useless and you must install qt for vlc,telegram and you need gtk for gimp .. so if you compile all of them I don't know how long it will take

Thx. Your answers seem to strengthen the point NOT to mix pkg with ports. I'll go with that.

If it's for server you can use only ports cause programs are lightweight

I'm considering a RasPi 4 as a FreeBSD webserver.

Yes it's cool !! Go with ports

Tips use lighttpd is optimized

Aha. Last couple of weeks I found out through web research, that nginx is still fastest even on the smallest machine. But the configuration for php+mariadb is quite hard to accomplish, as ngnix comes with its 'proxy logic'. I tried to separate different file types within the webserver. Not easy...

Tips use lighttpd is optimized

Maybe useful for the one or other: check out this inspiring website. It explains how to setup a personal website with lighttpd on a RasPi (under RasPiOS, though): https://cheapskatesguide.org/

does pine64 is an option for fbsd?

does pine64 is an option for fbsd?

You might check on https://t.me/PINE64_News

You might check on https://t.me/PINE64_News

Subscribed, thanks 🤝

I'm considering a RasPi 4 as a FreeBSD webserver.

I'm doing exactly this, and it is really nice to work with!

I'm using a setup where the / filesystem is mounted read-only: /var/ is automatically mounted in RAM. Then I've got unionfs overlays mounting a usb thumb drive (expendable) over system directories:

#!/bin/sh set -e if [ -z "$1" ]; then echo >&2 "Usage $(basename "$0") directory" exit 1 fi if ! target_dir=$(readlink -fn "$1"); then echo >&2 "Does not exist: $1"; exit 1 fi ( cd "$target_dir" find . -type d -mindepth 1 -maxdepth 1 | grep -v var find ./var -type d -mindepth 1 -maxdepth 1 ) | while read -r path; do path=${path#./} [ -e "$target_dir/$path/.rw_mounted" ] || touch "$target_dir/$path/.rw_mounted" if [ -e "/$path/.rw_mounted" ]; then echo >&2 "/$path already mounted, skip!" elif ! mount_unionfs /mnt/$path /$path; then echo >&2 "failed mount_unionfs $path /$path, skip!" fi done # force time synchronization ntpd -qg service ntpd start # create non-existing log files newsyslog -vC

☝️ This one I can run after bootstrap to turn it from vanilla FreeBSD to the configured system

Subscribed, thanks 🤝

Your answer is YES: FreeBSD does support Pine64. See: https://www.freebsd.org/where.html

I'm doing exactly this, and it is really nice to work with!

Thank you very much, Will. This is encourageing. I assume, you are using the official FreeBSD ARM image for RasPi3 and modify it, right? [https://wiki.freebsd.org/arm/Raspberry%20Pi#FreeBSD.2FARM_on_the_Raspberry_Pi_family]

Yes, the regular arm image. I'd really like one day to have my own flavour of the image, but I'm not expert enough with FreeBSD (I'm native from Linux). The regular image was installed, then I just changed /etc/fstab to have a read-only root. Then you reboot, and you're set to go.

The whole reason why I avoid having read-write root is that the SD cards thend to wear off fast, if for example you store log files on it

I own 9000 usb disks, while I don't want to buy more SD cards.

so I consider the USB disk expendable.

Obviously I've got a (tarsnap based) backup system that backs up the configuration files that live on the USB drive.

On a system like this I can basically fool around and explore. If I misconfigure I can reboot and it's back to normal. If the hardware fails, I can restore it in a second.

The whole reason why I avoid having read-write root is that the SD cards thend to wear off fast, if for example you store log files on it

All Sandisk SD-Cards are using wear-leveling algorithms since a couple of years now. I've been checking on that 2yrs ago. My strategy here was to use a very big SD-Card with a small installation (let's say approx. 10-20% of the card size). This should be relatively 'safe' in terms of endurance. It is hard to dig out empirical reports from the net, though...

I own 9000 usb disks, while I don't want to buy more SD cards.

Are you talking about a room full of disks!?? How does your data-management look like??? 😉

All Sandisk SD-Cards are using wear-leveling algorithms since a couple of years now. I've been checking on that 2yrs ago. My strategy here was to use a very big SD-Card with a small installation (let's say approx. 10-20% of the card size). This should be relatively 'safe' in terms of endurance. It is hard to dig out empirical reports from the net, though...

I see. Then it might be overkill, I don't know :)

Are you talking about a room full of disks!?? How does your data-management look like??? 😉

Nah, it wasn't a literal 9000.

😂

I just say that my installation runs on *one* USB disk. If that dies, I can remove it, toss it in the trash bin, insert the next usb disk, restore.

Since I've got many USB disks, I dont' care

Understood. That makes much sense.

Aha. Last couple of weeks I found out through web research, that nginx is still fastest even on the smallest machine. But the configuration for php+mariadb is quite hard to accomplish, as ngnix comes with its 'proxy logic'. I tried to separate different file types within the webserver. Not easy...

You can have nginx + php + mariadb .. But if you want to do an API server than you must pay for it on Nginx

I've also been considering tying up a USB-HDD with this RasPi4. It is definitely a very good choice, but it makes the question of a proper housing solution for the SOC more difficult.

You can have nginx + php + mariadb .. But if you want to do an API server than you must pay for it on Nginx

Thx for this. I was not aware of it.

Aha. Last couple of weeks I found out through web research, that nginx is still fastest even on the smallest machine. But the configuration for php+mariadb is quite hard to accomplish, as ngnix comes with its 'proxy logic'. I tried to separate different file types within the webserver. Not easy...

It's not that hard. You must run php-fpm on localhost:9000 f.e. and proxy access to *.php to localhost:9000

It's not that hard. You must run php-fpm on localhost:9000 f.e. and proxy access to *.php to localhost:9000

Yes. I'm using php-fpm. Problem was and still is to configure nginx for separate file types in /var/www, e.g. with subdirectories like <html>, <pics>, <php> , <css> and so on. So far, I was not able to alter the php-path from /var/www to /var/www/php

I was able to configure it running different static websites on the same raspi. But separating php-code from html-code is a bit trickier, as it seems.

Currently, I get either "No input file specified." or "404 Not Found", when I try to display /var/www/php/index.php in a browser from a remote host.

I've also been considering tying up a USB-HDD with this RasPi4. It is definitely a very good choice, but it makes the question of a proper housing solution for the SOC more difficult.

by USB-HDD you mean a proper disk, not a USB thumb drive, right?

'cause in that case, no, you don't have a SOC. :D You've a SOC with a KNOB

'cause in that case, no, you don't have a SOC. :D You've a SOC with a KNOB

a KNOB? 😊

…a literal knob, not an acronym

:D

😂

But hell yeah, I'd like to use my raspberry to stream stuff to $relatives. Like that proverbial movie that I obtained very legally. <COFF COFF> Or simply family photos, really

yes - bulky then.... I meanwhile have so much IT stuff here, that 'cabling' alone is hell... 😉

But hell yeah, I'd like to use my raspberry to stream stuff to $relatives. Like that proverbial movie that I obtained very legally. <COFF COFF> Or simply family photos, really

A FreeBSD (intranet-) webserver was a great digital diary. For external communication, I was still looking for people interested in freedombox on raspi.

But TG freedombox is pretty much dead since it came into existence. I already considered establishing a group of my own for this.... but I have no clue how much time effort this might be...

But TG freedombox is pretty much dead since it came into existence. I already considered establishing a group of my own for this.... but I have no clue how much time effort this might be...

Do you want to make a freedombox on a FreeBSD basis?

It's not that hard. You must run php-fpm on localhost:9000 f.e. and proxy access to *.php to localhost:9000

unix socket file is better, imo