**** BEGIN LOGGING AT Sun Jan 10 03:00:21 2021 Jan 10 21:13:51 Does the beagle bone no longer run the Apache2 server? Jan 10 21:14:18 it switched to nginx, pretty sure a long time ago Jan 10 21:15:16 i think it still has it at 8080. not sure. nginx proxy is running up front, for sure. Jan 10 21:16:08 what, why would you run both? o.O Jan 10 21:16:22 I don't see an apache package installed Jan 10 21:16:49 Okay, that makes sense. I was reading Molloy's book and it says that an apache server run by default upon installation Jan 10 21:17:05 I looked in /etc and saw an apache2 folder, but no services for it Jan 10 21:17:09 hmm, not here. looking for the rules. Jan 10 21:17:30 Zalymo: isn't molloy's book really dated? Jan 10 21:17:48 if the package isn’t installed, there’s your answer. Jan 10 21:17:49 (like books about computer-related topics tends to do really fast) Jan 10 21:17:54 version1 is, but version2 is more recent (published in 2019 I believe) Jan 10 21:18:10 I'm reading his "exploring beaglebone" book Jan 10 21:18:21 then the question is, how well has it been updated for v2? :P Jan 10 21:19:07 https://github.com/beagleboard/image-builder/commits/master/configs/bb.org-debian-buster-iot-v5.4.conf Jan 10 21:26:18 @rcn-ee where are the custom nginx rules again? Jan 10 21:36:26 nginx is whats running on my bbb based on all the images ive downloaded and checked sofar Jan 10 21:36:41 the setup stuff is in the /etc/nginx folder Jan 10 21:38:04 you can repopulate the apache2 folder tho and change over to it by adusting the init.d entries Jan 10 21:41:54 seems to me i recall a script in the tweak-tool area of the /opt directory i think that let you kill the bonescript and reset the httpd services back to port 80 as well Jan 10 21:42:56 no idea of it works tho as i just rebuilt without any of those packages and tossed all the nginx stuff and install apache2 Jan 10 22:15:44 why would you want to do that? Jan 10 22:18:19 just old and stuck in the mud and used it since it replaced cern Jan 10 22:18:21 lol Jan 10 22:18:31 lol fair enough Jan 10 22:18:46 to many years invested in something thats as relavent today as it was back then Jan 10 22:18:49 ;) Jan 10 22:20:14 to be honest in linux in general is like that for me as if i had to stop and relearn everthing the way things are currently done i might as well stop playing as all my time would be spent learning new things Jan 10 22:20:37 tho i will admit ive really heard nothing bad about nginx Jan 10 22:20:58 btw you know the init.d files aren't used anymore right? ;) Jan 10 22:21:15 if i had my way everyone would still be running slackware or freebsd Jan 10 22:21:37 (unless you also replaced systemd by sysv-init or whatever) Jan 10 22:21:38 ya i kinda suspected systemd took most of that over Jan 10 22:22:31 you'll find the service file in /etc/systemd/system/ (in the form of a symlink to the package-provided service file in /lib/systemd/system/ ) Jan 10 22:22:55 actually i just about got a debian 10 version running sysV Jan 10 22:23:27 just playing around trying to learn the ins and out of the image-builder projects Jan 10 22:24:19 normally i just build by hand or buildroot root but seen the image-builder gits floating around and have spent the weekend playing with them trying to understand more about how they all work Jan 10 22:24:22 my own code at work frequently has a hard dependency on systemd Jan 10 22:25:13 im more old school in my linux thinking but definately can see why systemd has come as far as it has Jan 10 22:25:15 it's super useful Jan 10 22:25:25 also, I care about boot time Jan 10 22:25:45 i just find it gets in my way as i need to stop and think about it all the time Jan 10 22:26:04 I once spent a little bit of effort on boot time optimization on a beaglebone and got this: https://liktaanjeneus.nl/boot.svg Jan 10 22:26:08 ya for sure if ones worried about boot times i can definately see why you would want the use of it Jan 10 22:26:22 zmatt: I like systemd - which is something that gets you flak wherever you go - but boot time is not on my short list Jan 10 22:26:26 (the graph is from systemd-analyze) Jan 10 22:26:36 cool Jan 10 22:27:14 philenotfound: yeah in our case it's integrated into an appliance that users will turn on and off so boot time is annoying Jan 10 22:27:15 ya thats pretty nifty Jan 10 22:27:58 systemd is presumably also a hard dependency for systemd-networkd, which I like a lot more than the alternatives I've seen Jan 10 22:28:27 zmatt: i'm still agreeing with the point you made :) Jan 10 22:28:32 someday someone needs to make a linux rootfs with all the old junk removed just to make things easier Jan 10 22:28:52 and I use its privilege-minimization features to containerize and harden our services Jan 10 22:29:07 seeing remnants of old things left make it kinda confusing for peeps that are more hobbiest Jan 10 22:29:24 or just old like me Jan 10 22:29:26 lol Jan 10 22:29:46 I've definitely done some of that cleanup on our systemd... like Jan 10 22:29:47 -bash: ifconfig: command not found Jan 10 22:30:26 oh, ok, i don't see containers as a security tool, more as a deployment tool Jan 10 22:30:36 it's both Jan 10 22:31:03 it's alos an important part of our update system indeed Jan 10 22:31:08 *also Jan 10 22:31:49 but I do like the fact that even if someone compromises one of the services they can get absolutely nowhere Jan 10 22:32:41 especially since one of our services is a roon endpoint (https://roonlabs.com/), which is pretty much a remote code execution service with no authentication Jan 10 22:51:29 can anyone help me figure out why I get this error? Jan 10 22:51:34 debian@beaglebone:~$ config-pin P8_08 hi Jan 10 22:51:35 ERROR: write() to /sys/devices/platform/ocp/ocp:P8_08_pinmux/state failed, No such device Jan 10 22:51:59 hmm Jan 10 22:52:08 are you booting from sd card or from eMMC ? Jan 10 22:52:27 oh wait Jan 10 22:52:28 never mind Jan 10 22:52:36 "hi" is not a valid pinmux mode Jan 10 22:53:26 so, this works, debian@beaglebone:~$ config-pin P8_08 gpio Jan 10 22:53:26 Current mode for P8_08 is: gpio Jan 10 22:53:36 I recently learned some older version of config-pin (written as shell script instead of in C) also supported some amount of gpio setup, even though that's something completely unrelated to pinmux setup Jan 10 22:54:12 though generally applications that use gpio would just set up the gpio themselves Jan 10 22:54:40 right, "gpio" is a pinmux mode Jan 10 22:54:56 though usually you want "default" rather than "gpio" Jan 10 22:55:10 yeh, that works too Jan 10 22:55:18 because weirdly the "gpio" mode of cape-universal means "gpio with internal pull-up/down disabled" Jan 10 22:55:33 which is something you should usually not do Jan 10 22:56:17 gpio_pu and gpio_pd modes are available to explicitly select internal pull-up or internal pull-down, while "default" uses whatever is the reset default pull direction for that pin Jan 10 22:56:19 so this page https://github.com/beagleboard/bb.org-overlays/tree/master/tools/beaglebone-universal-io Jan 10 22:56:39 shows hi is a valid mode Jan 10 22:57:03 yeah that's probably the old version, Jan 10 22:57:16 the fact the readme talks about 3.x kernels is a good hint that it's ancient ;P Jan 10 22:57:26 oh, that might explain it, I am porting from an old image Jan 10 22:57:40 bone_capemgr is also no longer used Jan 10 22:57:42 lol, thanks Jan 10 22:58:12 right Jan 10 22:58:59 is the config-pin utlity obsolete? Jan 10 22:59:04 note that all config-pin does is write the state name to a sysfs attribute, so it's pretty easy to do in any language, e.g. this is a python version: https://pastebin.com/MKtWJ8G8 Jan 10 22:59:11 definitely not Jan 10 22:59:25 though the old shell script has been replaced by a C version Jan 10 22:59:38 (a pretty horrible C version when I glanced at it :P ) Jan 10 22:59:52 whats the new way of doing "config-pin P8_08 hi" then? Jan 10 23:01:31 not sure if there's a nice util for setting a gpio (since usually that's done by applications rather than by shell scripts), I generally just write to sysfs directly... e.g. echo hi >/sys/class/gpio/gpio67/direction Jan 10 23:01:58 though I have an udev rule that creates convenient symlinks, which would make it more readably "echo hi >/dev/gpio/P8_08/direction" Jan 10 23:02:12 this is assuming the intent is to change the gpio to output mode, initially high Jan 10 23:03:37 (the "direction" attribute reads as "in" or "out", but when writing it to change direction to output it also needs to know what the initial output level needs to be, hence you write "hi" or "lo" instead of "out") Jan 10 23:04:02 or "high"/"low", can't remember what it wants exactly.. maybe it accepts either Jan 10 23:04:41 yeah, "high"/"low" (with "out" accepted as synonym for "low") Jan 10 23:06:08 i was using "config-pin" in a node.js module that executed the command Jan 10 23:06:31 that seems quite silly, use some nodejs module for gpios instead Jan 10 23:07:07 or bonescript, or even just plain nodejs code to write the sysfs attributes Jan 10 23:07:41 probably, I forget why I did it that way its been a while since I wrote this thing. Jan 10 23:38:56 I am using "onoff",  but my notes say that module doesn't support setting the pullups on the beaglebone. So I wrote a little helper for onoff to set the pullups using config-pin. Jan 11 00:01:05 anyway, I'll just use gpio_pu/pd for now to set the input resistors, just adds and extra line of code. Jan 11 00:01:16 thanks @zmatt! **** ENDING LOGGING AT Mon Jan 11 03:00:17 2021