ratos 2.1-rc2 issues (no initial config)
I created a fresh SD card using the July 28 ratos 2.1.0-rc2 image for rpi.
On first boot, I got a "unknown config object 'bed mesh'" error.
I attempted to update all the repos to the latest, starting with the ratos configurator, and now the ratos repo is marked invalid/uninitialized by the management tool:
and also I get the error:
Existing command 'TEST_RESONANCES' not found in RatOS override
and the initial config tool never comes up. am I missing something totally obvious that everyone should be remembering, like I was supposed to go to ratos/initial-config instead of just ratos or something?
56 Replies
https://os.ratrig.com/docs/installation#updating-ratos is the latest instructions for updating
Installing RatOS | RatOS
This is preliminary documentation for the upcoming v2.1.0 release. The work is still ongoing and the documentation is not yet complete.
aha, i see there's a /config or /configure url (which is it?) to use.
i guess you do /configure then /config? that's what it looks like
On first boot, I got a "unknown config object 'bed mesh'" error.That was fixed a long time ago, generally errors before updating an old RC can be ignored.
I attempted to update all the repos to the latest, starting with the ratos configurator, and now the ratos repo is marked invalid/uninitialized by the management tool: and also I get the error: Existing command 'TEST_RESONANCES' not found in RatOS overrideThis is normal (follow the process in the docs). It's also documented in the mainsail notification as well as the PSA over in #announcements btw Two different things. Mainsail Machine panel = /config RatOS configurator = /configure
lotsa different places to look
(yeah it's confusing that mainsail calls the machine tab
/config)
same message in all of them
Just stick to the docs.
(i was simply trying to relay that i did my best to tell you about this update)I appreciate it, may i suggest a pinned message with the appropriate links to the releases and the docs, and maybe also have the links to the docs from the github releases? Probably would have made it easier for my situation at least - since I'm not starting a brand new printer build from scratch, finding the links in the build docs is kind of troublesome
(is there such a pinned message already? I did look, fwiw)
There are two steps in this latest one as the configuration was merged into the configurator which needed significant FS and moonraker config changes. New moonraker config changes aren't picked up until moonraker is restarted, and the new pinned_commits aren't picked up till the update manager is refreshed after moonraker has been restarted.
pinned messages are almost never read, but i can certainly link to the docs from the GH releases!
Also go to #role-assignment and click the ratos topic 🙂
I do sometimes remember to look in pinned messages but not always, I freely admit
thought I had... checks
it's not in your roles, so you won't get pinged when stuff like this needs a bit of extra attention (which is rare, especially on final releases, but it does happen)
weird, i have it checked.
along with several others
Oh i didn't expand the list
that's my bad
at this point do you recommend I start with a fresh sd card again, or just continue on my clumsy way? 🙂
You're almost there, just restart moonraker, refresh the update manager and update the remaining packages.
Except @za_naimis, did you click "update all"?
Because then you broke it and would probably be better off with a fresh install, following the update instructions.
You can recover it, but it's a bit of a mess.
Moonraker has an interesting "feature" where it just completely ignores the git repository and forces a merge using the values from the loaded configuration (ie, it does a
git pull origin [configured branch]). So you end up merging the old configuration repository into the configurator deployment branch, it's ... not great.i updated the configurator then clicked update all.
it was a fresh install to begin with, so it's not really a burden to do it again
i updated the configurator then clicked update all.Yeah that broke it
ah well 😄
It's annoying. There was no mitigation other than instructions.
the hardest part of this is getting the stupid sd card out and putting it back in the rpi on the standard vc3 electronics panel
fwiw, i would probably have done what Moonraker does the same way if i had written it myself, thinking it was "safer" (it's not though). RatOS is "a bit" of an edge case as well so can't say i blame them :D.
Yeah it's not in a great position
longer standoffs help
i should probably do that. I have some nylon ones I bought in a kit when I added a proper fan to the octopus.
Oh btw, there is a way you can fix this broken state without reflashing:
1) Restart moonraker
2) Refresh update manager
3) Run hard recovery on the "invalid" packages.
4)
will that work even if they appear to be uninitialized, as ratos was?
You picture says "invalid". I'm not sure there's an "uninitialized" state. Unless you mean the starting warning about refreshing, before refreshing the update manager on first run?
uninitialized is what was being reported by git. I had captured a screenshot of that but didn't save it, alas
in what directory?
"RatOS"
~/printer_data/config/RatOS ?I believe so, but I can't recall if it gave the whole path, all I remember is that it was "RatOS"
You would've had to navigate to it via ssh manually
we'll see if any of this persists after I actually follow the instructions 🙂
👍
Rat-OS 3.0!
looks like a string indexing issue: (one extra character in one of the version signatures)
Yeah... painlessly upgrading from 2.0 seems to get progressively unrealistic.
Old bug, fixed ages ago, you forgot to update (ie, follow the instructions).
Just realized the heading there is stupid (left over from old docs).
Yeah at this point might as well say it's a major release
the instructions say to run through the configure process before updating
No it does not
see the screenshot right below the commend you replied to
yeah that's what I've been working against. In the "wifi or ethernet" section, it tells you to follow the /configure url and process, which is where I ran into that octopus firmware id bug, followed by "preparing your control board" and "updating ratos"
unless the /configure URL does something different when connected over wifi. But when I followed that URL, it ran through the configurator. I had the option of setting up wifi but i never use it so I skipped that step. but it does say:
When connected, open http://RatOS.local/configure or http://192.168.50.1/configure and follow the steps presented there to setup wifi and your printers hostname.but does not say to stop after doing wifi setup, or anything at all about skipping the wifi setup
back at this again. It's not clear if "refresh the update panel" was meaning to refresh the web page or to click the "Check for updates" button. I did the latter, followed by the former. I'm going to try continuing with the instructions and do the update all. Maybe the pictured issue is due to ratos and configurator being combined? That seemed to be implied by the instructions.
and i did update ratos first followed by configurator
went away after i manually updated everything and the ratos: core ... repo showed up.
It tells you the specific steps to follow though, and also says very specifically what to do once it's booted. I'll try and make it more explicit.
Meant to say "refresh the update manager". Update all will break everything at this point, don't do that.
it looks like moonraker hasn't been manually restarted here
we're obviously not seeing the same thing. What I described is how things progressed when I followed the directions as they were written
"Wifi setup wizard". What you went through is a whole lot more than wifi. Nevermind, i will explicitly tell users to stop after the wifi configuration and the system has rebooted.
You might have just needed to refresh your browser in this case.
After restarting moonraker that is
It probably fixed itself when you updated moonraker.
I would suggest making it clearer at least for those of us who skip the wifi setup. I suspect that the steps are very different in that case.
It's the same step, you just click "skip" instead of "save", and thus there's no reboot. It's still the wifi step.
anyway.. like i said i'll add an explicit instruction to stop.
k
This is only temporary btw, it'll do this automatically in the future.
it seems to be alive at the moment, unless you think that the way i got to this point is flawed and I need to do it over again?
granted I haven't tried setting up any of the details like webcams yet
or, you know, trying to print something 🙂
No it's ok, it would just be nice to know what went wrong so i can't prevent the next guy from being confused.
This is barely related to RatOS anyway (it's all crowsnest/mainsail) 🙂
Everything should be fine now.
I'm bridging between this and other related issues, but after configuring both of my logitech webcams and running a home and tilt, I wasn't seeing dropped frames. I'll probably start adding back my customizations bit by bit until I start seeing dropped frames again.
3DO camera is probably toast, so I'm unlikely to reconfigure it, and I'm currently waiting on a new one along with all the other toolhead parts
one of the most recent changes I made was adding the
[exclude_object] section.
This makes me very suspicious given what I've heard about it having a significant impact on printer performance.Not
[exclude_object], enable_object_processing in moonraker's file manager (which is required to use [exclude_object]) is the big oneI don't think I ever explicitly did anything with enable_object_processing, unless I made a change and immediately forgot about it. All I remember doing is the exclude_object section and the slicer settings
though maybe that's why the one time I tried to cancel an object, it tried to print it anyway. That is, the mainsail interface seemed to respond and work as it was expected to, but ... the underlying print cancel didn't work
looked for enable_object_processing in every file under /home/pi and other than documentation (*.md) it was set to false in both the old and the fresh ratos installs.
I had the invalid packages also but what I found out was that when waiting a bit longer then the browser refreshed and all where able to update. After that I had 116 system packages not wanting to install but after a few tries they did and everything worked. So it looked like extraordinary 😂 paitience was key in my case.
I also had the bed_mesh error and the resonance error, but that was fixed after updating everything and set the adxl for x and y..
please make your own post