What is the problem?
Recently I cannot connect to two different Raspberry PI instances of Octoprint. One is an RPi 3+, the other is an RPi 4+. Both are on 1.4.0. When I try to connect using their respective IPs, I get "Loading Failed -- Socket connection failed .." Both instances are fine when accessed from a browser on an android tablet or from the android Printoid app on my phone.
What did you already try to solve it?
rebooting both RPI instances.
Attempted access from Chrome and Firefox browsers
Logs (octoprint.log, serial.log or output on terminal tab at a minimum, browser error console if UI issue ... no logs, no support!)
Chrome browser log shows (I swapped in SERVER ADDRESS):
Starting dependency resolution...
packed_plugins.js?846f05c7:2123 SlicerSettingsTabViewModel
packed_core.js?3514a079:17074 ... dependency resolution done
packed_core.js?3514a079:17384 Initial application setup done, connecting to server...
packed_core.js?3514a079:8263 Loading error: Socket connection failed. Please check prior messages and 'octoprint.log' for possible reasons.
UiStateViewModel.self.showLoadingError @ packed_core.js?3514a079:8263
packed_plugins.js?846f05c7:2155 undefined
packed_core.js?3514a079:15010 Connected to the server
htmlfile?c=_jp.aqn1oe5:8 Uncaught DOMException: Blocked a frame with origin "http://*SERVER ADDRSS" from accessing a cross-origin frame.
at http://SERVER ADDRESS/sockjs/673/fzjatkuh/htmlfile?c=_jp.aqn1oe5:8:20
htmlfile?c=_jp.aqn1oe5:10 Uncaught TypeError: Cannot read property 'message' of undefined
at p (htmlfile?c=_jp.aqn1oe5:10)
at htmlfile?c=_jp.aqn1oe5:15
htmlfile?c=_jp.aqn1oe5:10 Uncaught TypeError: Cannot read property 'message' of undefined
at p (htmlfile?c=_jp.aqn1oe5:10)
at htmlfile?c=_jp.aqn1oe5:18
DevTools failed to parse SourceMap: chrome-extension://gighmmpiobklfepjocnamgkkbiglidom/include.preload.js.map
DevTools failed to parse SourceMap: chrome-extension://hdokiejnpimakedhajhdlcegeplioahd/sourcemaps/onloadwff.js.map
DevTools failed to parse SourceMap: chrome-extension://gighmmpiobklfepjocnamgkkbiglidom/include.postload.js.map
Additional information about your setup (OctoPrint version, OctoPi version, printer, firmware, browser, operating system, ... as much data as possible)
Octoprint 1.4.0
Windows 10 1909
Boy, that was an excellent suggestion. Unfortunately no luck. I put in an exception for the address and reloaded the page, but no luck. I also went over to Edge (where I don't have an ad blocker) and that threw the error as well.
Thanks, greatly appreciate the help. I'm on the octopi server using winscp, but can't for the life of me find where octoprint is actually installed, hence, can't find the octoprint log files.
From within WinSCP, no variation of: ~/.octoprint/logs appears to generate a valid directory location.
This is probably something really simple, but it is escaping me.
Found it. home/pi/.octoprint and that got me to the logs.
Right now the printer is helping support printing Prusa face shields for Covid-19. I'm looking at the log file, but haven't had a chance to restart the server in safe mode.
here is what I see when trying to log in from my PC:
2020-03-29 06:38:19,894 - octoprint.server.util.flask - INFO - Passively logging in user 3DMaster from ::ffff:10.0.0.28
2020-03-29 06:38:19,903 - octoprint.access.users - INFO - Cleaning up user session 320071E51FD24FCCB2D99A537508DE3D for user _api
2020-03-29 06:38:19,911 - octoprint.access.users - INFO - Logged out user: _api
2020-03-29 06:38:19,914 - octoprint.access.users - INFO - Cleaning up user session 73B6916859DA4090AD017C2093140352 for user 3DMaster
2020-03-29 06:38:19,919 - octoprint.access.users - INFO - Logged out user: 3DMaster
2020-03-29 06:38:19,923 - octoprint.access.users - INFO - Cleaning up user session 7C4E175BD5E14A6B9C73BD0FB4E4C7FC for user _api
2020-03-29 06:38:19,953 - octoprint.access.users - INFO - Logged out user: _api
2020-03-29 06:38:19,956 - octoprint.access.users - INFO - Cleaning up user session 47114A8EBA6548ED92A3D1B7619AC167 for user _api
2020-03-29 06:38:19,958 - octoprint.access.users - INFO - Logged out user: _api
2020-03-29 06:38:19,965 - octoprint.access.users - INFO - Logged in user: 3DMaster
2020-03-29 06:38:21,640 - octoprint.plugins.Octoslack - ERROR - Slack RTM message processing error: coercing to Unicode: need string or buffer, NoneType found
2020-03-29 06:38:27,367 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:28,021 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:28,614 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:29,220 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:29,829 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:30,522 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:31,124 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:31,701 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:32,399 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:38,775 - octoprint.server.util.sockjs - INFO - Client connection closed: ::ffff:10.0.0.28
2020-03-29 06:38:40,742 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:41,536 - octoprint.server.util.sockjs - INFO - New connection from client: ::ffff:10.0.0.28
2020-03-29 06:38:47,760 - octoprint.server.util.sockjs - INFO - Client connection closed: ::ffff:10.0.0.28
Looks like It's using IPV6 which Android and Linux fully support. Windows has issues with IPV6. may want to add ipv6.disable=1 to the cmdline.txt file in /boot and see what happens
In a case like this, I might put the microSD into an adapter and then plug it into my MacBook. I'd then jump to a terminal and... sudo nano /Volumes/boot/cmdline.txt to edit it directly, then Eject and boot the Pi like that.
But in your case, you'd need to use PuTTY to get a session to the Pi, then sudo nano /boot/cmdline.txt.
Well this is really strange or should I say sporadic. I deactivated IPV6 for the one Octopi instance and rebooted. At that point, the instance was not accessible at all, regardless of platform. Went back in and munged up the undo to the point I just had to recreate the original cmdline.txt by hand (couldn't upload a new one due to permissions). After that and a reboot, the instance is now available and I'm using it. The other instance seems to be working as well (I had a working browsers session for that one that I had been trying to protect, but wound up needing to reopen it and it seems to be working as well). As I get some free time, I'll try to see if there is a pattern.
Sorry for the delay. It is definitely inconsistent. Octopi Instance 1 is currently inaccessible, but it has always had a DHCP reservation. Octopi Instance 2 is currently accessible, did not originally have a DHCP reservation, but I did add one a day or so ago.
I have no idea what the underlying issue is. I'm trying to get Instance 1 to restart in safe mode but haven't had the time to dig deep, and when I try to run the cmd to set safe mode, it says something to the effect "bash: octoprint: command not found". My linux-fu is lacking.
I'm in the /home/pi/oprint/bin directory and see the octoprint file
If you haven't changed the hostname on one of the RPi images, they are both trying to identify themselves as "octopi" which will certainly confuse any DHCP server running on the your network.
From a command line on one of the RPi's run "sudo raspi-config" and under the Network Options, change the hostname to "octopi1". If you can get to the other one, change the hostname to "octopi2". This frees the hostname "octopi" for use when you buy that third printer
The system running your DHCP server will probably need to be rebooted. If you can configure your DHCP server, then removing any old reservations for these two RPi systems might be useful. Its likely that the system with the DHCP server is also providing a local DNS server as well and if so, DNS may need to be cleaned up. Lastly, systems which had trouble connecting may need to be rebooted (or have their DNS cache flushed).
BTW, all of the RPi systems (each with unique names) on my LAN have both IPV4 and IPV6 addresses. They get their IPV4 addresses via DHCP. I don't think you need to limit your RPi systems to IPV4 only.
Quick update - The instance names are / were unique, but I went ahead and gave Instance 1 a new name anyway - cr10spro-v1. It was again inaccessible this morning. I tried deactivating my BitDefender security app, but that didn't have any beneficial affect.
I did power cycle cr10spro-v1 and that did enable me to get into the instance.
I have not been able to cycle my Orbi - that will have to wait for this evening. Everyone home, running school remotely and my son has a virtual interview this afternoon as well. Don't want to jinx anything.
I own the Orbi pair and they work great for me. I will note that the Avahi service in tandem with Netgear routers (Orbi) or similar will negotiate hostnames sometimes in ways that are weird.
cat /var/log/syslog | grep Avahi
Sometimes this will say, essentially, something like "I know you just asked for octopi as a name but I'm going to issue you octopi-2 instead. This typically happens right after you pull a microSD card from one Pi and place it into another. The 2nd Pi's Avahi service will come up, ask if it's okay to broadcast as octopi, the Netgear router will say "no" and then Avahi will decorate its name to octopi-2, ask the Netgear, it says "yes" and then it's now this.
Don't want to jinx anything.