<Guest48>
What is the operating temperature range of the entire beagleboard AI?
Guest48 has quit [Quit: Client closed]
ikarso has joined #beagle
florian has joined #beagle
lucas_ has quit [Quit: Leaving]
lucascastro has joined #beagle
av500 has joined #beagle
Shadyman has quit [Quit: Leaving.]
philenotfound has joined #beagle
buzzmarshall has joined #beagle
JustBoB has joined #beagle
<JustBoB>
Hey folks. Just a question. Is there any GUI based image available for AI Beagle? Checking here and all of them are without GUI. The original one which came with the board is with explored certs and kina brick. https://beagleboard.org/latest-images
<JustBoB>
overall im finding image support being bit obsolete. What people do in general?
<JustBoB>
Thanks. Would be good to update webpage too with some of these :)
<JustBoB>
by any change cloud-inti is supported for these imageS?
<rcn-ee>
JustBoB, sorry never heard of that... so not pre-installed... looks interesting..
<rcn-ee>
ugh, it pulls in eatmydata... oh does that globally disable ext4/fsync, or do you have to force it..
<rcn-ee>
okay, you need to call it from the wrapper.. so should be safe..
<JustBoB>
We use it everywhere to bootstrap the devices and virtual machines. quite neat to bootstrap multiple devices. I don't think It can be called from wrapper as it very early in the process of OS boot
<rcn-ee>
JustBoB, biggest issue, we have a few startup scripts that assume, uid 1000 or user debian, so need to massage it...
<JustBoB>
all this is before that. All cloud providers uses them to bootstrap vms. I suspect all your script could go into cloud-init (no-cloud provider) and just do all the job there. Ok, gonna try to create issue with as much details as I can
<rcn-ee>
JustBoB, from command line, you can either run: sudo beagle-flasher
<rcn-ee>
and it'll start coping the microSD to eMMC
<rcn-ee>
or: sudo enable-beagle-flasher and it'll do it on the next "reboot"..
<rcn-ee>
if you have a ton of services setup, the "reboot" flasher works better.. (it's just where rsync runs..)
<rcn-ee>
as base image can be done either way..
<zmatt>
beagle-flasher doesn't first stop all services not needed for flashing or ssh ?
<rcn-ee>
JustBoB, didn't you say the BBAI? use "sudo cp -v /etc/beagle-flasher/bbai-microsd-to-emmc /etc/default/beagle-flasher"...
<JustBoB>
I dont have much. I basically need Synpse agent (https://synpse.net/ ) and docker. All other management comes from Synpse and is deployed as containers. Im trying to write a blog post how to use BB and Synpse
CrazyEddy has joined #beagle
<rcn-ee>
zmatt, correct it doesn't stop anything.. so not 100% safe..
<JustBoB>
yap, sorry, copied from command from forum
<rcn-ee>
JustBoB, just double checking.. it wouldn't boot. ;)
<rcn-ee>
zmatt, main change... user doesn't have to figure out how to use nano or sed.. i'm working on moving it back to buster..
<zmatt>
rcn-ee: should be easy enough to fix... just create a .target that declares dependencies on the things you want to keep running, give it AllowIsolate=true, and then use systemctl isolate to stop everything that's not a dependency of that target
<JustBoB>
`Debian 11.x (Bullseye) Xfce Desktop Snapshot` should this after flashing boots to X env? after flashing I'm getting no Xfce
<rcn-ee>
JustBoB, over hdmi you should see xfce load, there is a massive wait right now on first boot... not sure what's casuing the delay yet..
<rcn-ee>
it's using the etnaviv open source 2d driver.. but yeah it's a massive regression from Buster..
<rcn-ee>
right now that's just cockpit-motd.service, working on killing that too..
<zmatt>
I'm checking how wait-online works exactly
<rcn-ee>
my biggest worry, PocketBeagle when no ethernet is connected..
<zmatt>
because the intention *is* that network-online.target succeeds when you actually have network up, i.e. it's means to be pulled in by services which _require_ working network
<zmatt>
the beaglebone should not have any such service by default
<zmatt>
but it's good form to ensure it does work for people who do need it
<zmatt>
ah, setting all of them RequiredForOnline=false is what breaks --any
<zmatt>
since RequiredForOnline=false means the interfaces are ignored, so then it never finds an interface up
<rcn-ee>
so 'one' needs to be enabled... ugh... PocketBeagle, BeagleBone, Wireless..
<rcn-ee>
we never have 'one' always...
<zmatt>
no, just enable them all, but use --any
<zmatt>
and by "all" I mean eth0 and wlan0 ... I don't think what the usb interfaces provide is what would generally by meant by "being online"
<zmatt>
*be meant by
<zmatt>
again, nothing on the image should pull in network-online.target by default, but e.g. if someone writes a service that requires internet access and doesn't want to have to deal with dynamic network changes, that's where you'd add a dependency on network-online.target
<rcn-ee>
so with eth0 and wlan0... 7.411s systemd-networkd-wait-online.service
<zmatt>
is that the time it takes for one of them to get link up? is it still being invoked with --any ?
<zmatt>
obviously anything that depends on network-online.target will not be able to start very early in the boot process
<zmatt>
that's one of the big downsides of using network-online.target ... but for some people that may be a price they're willing to pay to be able to assume the network is already up in their own code
<rcn-ee>
10.192s with -ignore=usb0 --ignore=usb1 --any --timeout=10
<zmatt>
it should not have a timeout, that breaks things
<rcn-ee>
lost 2.5 sections by removing --interface=eth0 --interface=wlan0
<zmatt>
also, why are you timing your boot? again, nothing should be pulling in network-online.target on the bbb
<zmatt>
so the time it takes is unimportant
<zmatt>
since it should get run
<zmatt>
*shouldn't
<zmatt>
certainly nothing that actually stalls the boot process in a user-relevant way
<rcn-ee>
trying to figure out what's causing xfce to take forever to load... it's the only outlier https://paste.debian.net/1225714/
<zmatt>
rcn-ee: can you share the svg from systemd-analyze plot ?
<zmatt>
you said network-online.target was being pulled in by cockpit-motd.service ... whatever that is
<rcn-ee>
so yeah.. wonder if it's just entropy and xfce...
<zmatt>
so, don't put a low timeout on network-online.target ... by definition it should wait for network online, that's the purpose it provides to whichever unit is depending on it
<zmatt>
or any timeout really
<zmatt>
on systemd-networkd-wait-online.service I mean
<rcn-ee>
i'll stick it back to it's default. i'm not using it..
<zmatt>
the --any option sounds like a good idea, can't imagine why it's not default (assuming it isn't)
<rcn-ee>
it's not...
<zmatt>
and RequiredForOnline=false on the usb interfaces
<rcn-ee>
yeap, usb0/usb1 are too be locked out at all cost..
<zmatt>
yes, no need to use commandline args for that, RequiredForOnline=false in the [Link] section of the .network takes care of that
<rcn-ee>
makes you wonder thou.. is anyone shipping systemd-networkd 'enabled' by default... seems like some of the final tweaks are just missing (--any)...
<zmatt>
I'm still checking what exactly --any vs default does, since even by default it apparently only needs "at least one link to be online" though it needs all of them "fully configured or failed"
<zmatt>
my ethernet interface shows as "configured" even when not plugged in
<zmatt>
then again usb0 on a bbb here shows as "configuring"
<rcn-ee>
usb0 should change from configuring -> configured when it gets plugged in..
<zmatt>
hmm, so my ethernet when unplugged when networkd starts indeed gets stuck in "configuring", and in fact only goes to "configured" well after the operational state is "routable"
<zmatt>
but
<zmatt>
it subsequently remains "configured" even after unplugging
<zmatt>
I'm not sure what systemd-networkd means by this
set_ has quit [Ping timeout: 256 seconds]
vagrantc has joined #beagle
ikarso has quit [Quit: Connection closed for inactivity]
<zmatt>
I still don't quite get the logic behind the link state being "configuring" while carrier is down (but not going back to "configuring" if the carrier goes down again once configured) and wait-online failing on that if --any isn't specified
<zmatt>
it is possible to get the interface to go to "configured" even if the carrier is down by ConfigureWithoutCarrier=yes (in which case you'll also want IgnoreCarrierLoss=no) in the [Network] section but I'm not sure if that has bad side-effects
<rcn-ee>
thanks zmatt i think i have eerything cleaned up in git, and pushsing deb builds out..
set_ has joined #beagle
<CoffeeBreakfast>
I just bought one of the latest bbg available in my country ... and I haven't seen a bbb for sale in a long time. I have a bbb though..
<zmatt>
CoffeeBreakfast: farnell says they'll get 10,000 BBBs in stock in 2-3 weeks
<zmatt>
CoffeeBreakfast: digikey has 5337 BBBs in stock
<rcn-ee>
oh man! thanks @zmatt those have the new phy fix.. i need to pick a up couple asap..
<zmatt>
woo
set_ has quit [Remote host closed the connection]
<CoffeeBreakfast>
Import taxes are insane these days. Well.. at least I have a couple now..
set_ has joined #beagle
starblue has quit [Quit: WeeChat 3.0]
Guest95 has joined #beagle
<Guest95>
hi can anyone
<Guest95>
help me
<Guest95>
end Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)
<Guest95>
help please
<zmatt>
uhh
<zmatt>
how'd you manage that
<Guest95>
i don't get it
<Guest95>
what do u mean
<Guest95>
>
<Guest95>
?
<zmatt>
this needs a lot more context... what device are you using? what image are you using? what are you trying to do? what is the context of this error? can you share the whole kernel log? (please share using pastebin.com, don't paste it directly into chat)
<zmatt>
Guest95: do not send unsolicited private messages
<zmatt>
Guest95: also, the fragment you pasted is showing kernel 4.14.108-ti-r113 .. but you claimed you were using AM3358 Debian 10.3 2020-04-06 1GB which does _not_ use that kernel
<zmatt>
I can't really tell anything from this other than that the kernel didn't see any mmc devices
<Guest95>
how??
<Guest95>
when i restart it
<Guest95>
it can boot from mmc
<zmatt>
I have no way to guess based on the negligible information provided
<zmatt>
okay, so what are you doing to cause t to fail?
<Guest95>
what other info do u need
<Guest95>
?
<zmatt>
*cause it
<Guest95>
i rellay appreciate your help
<zmatt>
I have no idea what you're doing or trying to do, what steps are leading you to this error... all I know is that you're apparently trying to do something weird with u-boot, and that you're having trouble booting your kernel
<zmatt>
and that what little information you provided isn't actually correct (since this kernel isn't the one on the 10.3 2020-04-06 image you said you're using)
<zmatt>
I cannot magically guess what you might have done wrong if you don't even know what you did in the first place
<zmatt>
*if I don't even know
<zmatt>
rcn-ee: you any idea why the kernel might not see any mmc devices? (since it's not listing anything after "here are the available partitions:")
<Guest95>
can i send
<Guest95>
commands
<Guest95>
here
<rcn-ee>
Guest95, use pastebin for anything over a line..
<rcn-ee>
zmatt, without a full boot log (and quiet) disabled, it's a guess. .;)
<rcn-ee>
i'm afraid to ask them that..
<rcn-ee>
probally /dev/mmcblk1 -> /dev/mmcblk0 name issues..
<rcn-ee>
Guest95, you seem to be mixing a bunch of different guides... Which guide are you trying to do right now? igrning all other guides you'd touched?
<zmatt>
Guest95: honestly, the list of topics in this course doesn't sound like it's particularly suited for for novices
<zmatt>
like, details of the boo process, TFTP booting, buildroot, building kernel and bootloader, NFS booting... that's all really advanced stuff
<zmatt>
some of it is really obscure... like, serial booting? wtf?
<zmatt>
who on earth uses that
Guest95 has quit [Ping timeout: 256 seconds]
Guest95 has joined #beagle
<rcn-ee>
the comments on that course are hilarous... it's a 10 year old version of Angstrom... no way those are 5 stars...
Guest95 has quit [Ping timeout: 256 seconds]
<zmatt>
Guest95: regardless, if you can clearly communicate exactly what steps you performed then we may be able to provide feedback, but otherwise there's no way to guess
Guest95 has joined #beagle
Guest95100 has joined #beagle
<Guest95100>
?
<Guest95100>
i'm guest 95
<Guest95100>
sorry
<zmatt>
you know you could also try setting a better name than the "Guest" and some digits
Guest9512 has joined #beagle
Guest95 has quit [Ping timeout: 256 seconds]
Guest95100 has quit [Ping timeout: 256 seconds]
Guest9512 has quit [Ping timeout: 256 seconds]
buckket1 is now known as buckket
JustBoB has quit [Ping timeout: 256 seconds]
starblue has joined #beagle
<jfsimon1981>
Good evening,
<jfsimon1981>
the test lab is online
<jfsimon1981>
zmatt, the modbus probe indeed responds immediately, it does not wait 3.5 char ...
<jfsimon1981>
I asked the manufacturer. Someone suggested 3.5 char is a maximum recommended wait time, not mandatory minimum. I'm a bit lost but i can perform tests, great.
<set_>
Nice!
<set_>
Way to go, fine person.
<set_>
I am working on something completely different. Motors (still)!
<CoffeeBreakfast>
(to all) What is your favorite new command that you discovered the past year? (from someone on twitter). I go first: `remake`
<set_>
Twitter? Does twitter.com have commands now?
<set_>
I twitted a couple times already in the last month. But...I was unaware of their commands.
<set_>
I found macrofab outside of twitter for building my TIDA-00320!
<set_>
Servos!
<set_>
Oh and @zmatt: I saw your posts on the forum. Thank you for still pitching in to support people who are lost in a world of electronics...
* set_
<<<
<CoffeeBreakfast>
I mean Linux commands
<set_>
Oh. I learned from jfsimon1981 about the / -name command when searching for specific things in the file system.
<set_>
I am sad to admit, I know very little of Linux commands on Twitter b/c they will not allow me to have their SDK any longer.
<set_>
I am boycotting them.
<set_>
A lot of my source was dedicated to their SDK for use in the system of twittering. Now, they say no to me. Boo!
<set_>
"How are you going to use our software," they asked. I told them exactly how I would use it. Blam. "No set_, you cannot have our SDK any longer."
<set_>
CoffeeBreakfast: I understand your question. I was poking fun w/ you. Please forgive me.
<zmatt>
jfsimon1981: wtf, no it's most definitely the minimum _required_
<zmatt>
jfsimon1981: "In RTU mode, message frames are separated by a silent interval of _at least_ 3.5 character times." (the _at least_ is underlined in the spec itself)
<zmatt>
jfsimon1981: and the maximum idle time between two consecutive bytes of a single frame is max 1.5 character times
<zmatt>
jfsimon1981: it also explains that the 3.5 char idle is used to delineate frames: "When the link is in idle state, each transmitted character detected on the link is identified as the *start of a frame*. The link goes to the "active" state. Then, the *end of frame* is identified when no more character is transmitted on the link after the time interval t3,5."