I am also facing this issue, using the standard upgrade with osp-config.sh, I also manually attempted to migrate the database using the above commands because this has fixed these issues prior.
When the database is upgraded, you cannot view posted videos or login without the following error being thrown in /opt/osp/logs/osp-error.log
On migration to new hardware, trying to import the OSP backup file (76MB) is too large. I have disabled nginx client max body size, now it looks like it's a nginx -> flask issue. There is nothing in /opt/osp/logs/osp-error.log nor nginx. Is there a way to import the json file manually using flask or something?
I was able to import all tables by rsyncing the entirety of the mysql database, I will close this but I'm genuinely curious if there is a way to do it with command-line in the future
On migration to new hardware, trying to import the OSP backup file (76MB) is too large. I have disabled nginx client max body size, now it looks like it's a nginx -> flask issue. There is nothing in /opt/osp/logs/osp-error.log nor nginx. Is there a way to import the json file manually using flask or something?
that seems to have solved it. thank you for your time!
sure, here is a screenshot
running
sudo python3 manage.py db init
throws the following error
completing it with db upgrade throws the following errors
sorry to be a bother again, I managed to get it to work 3 weeks ago with your help but upgrading to the most current branch has caused SQL issues again and I've been racking my brain trying to figure out how to solve it -- current log pastebin is here
Database issue after upgrade to the latest commit, the 'single server' upgrade was ran, so was the 'core update' individually and also the 'database upgrade' https://pastebin.com/tkKaEd97
I ran "python3 manage.py db init" and ran into this failure https://pastebin.com/CYwdF0HX
looks like the database has an issue, would this be related to the password salt changing? https://pastebin.com/qTHvKb2Z
this error is thrown only when you try to delete a user -- modifying user roles are not effected however password recovery emails are also not sent for old entries, new entries they are (i.e. new user)
I spent a little bit reading logs and thinking, I completely terminate any HTTP request server side and it's forced to HTTPS with Cloudflare so switching the protocol in the auth_osp.py from HTTP to HTTPS resolved the live chat issue. Thanks for being a soundboard
I upgraded everything last night and have mostly ironed out the kinks, but nothing can be streamed because the ejabberd service is throwing errors. Originally it was throwing auth_osp.py errors because the installation (and reinstallation, and re-reinstallation) did not put the auth_osp.py file into the configuration directory. Manually moving it allows the ejabber config to start up but it then throws this error any time anyone tries to view a channel and provides no chat (just a spinning modal)
yeah changing it from 127.0.0.1 to localhost, the machines ip produces the same error. I installed the entire thing from scratch on a VPS just to see and the same issue persists on a fresh Ubuntu 20.04 installation with the prior database.
I lost my original database password salt in the upgrade, could that be the issue? I notice in the log it's erroring checking passwords. Is there a way to reset all user tokens for ejabber in the database?
I upgraded everything last night and have mostly ironed out the kinks, but nothing can be streamed because the ejabberd service is throwing errors. Originally it was throwing auth_osp.py errors because the installation (and reinstallation, and re-reinstallation) did not put the auth_osp.py file into the configuration directory. Manually moving it allows the ejabber config to start up but it then throws this error any time anyone tries to view a channel and provides no chat (just a spinning modal)
not sure why the database wasn't automatically altered as i've been incrementally upgrading. dropping that column and creating the column solved the issue, thank you.
when viewing a live stream that has not failed after server upgrade, vising "my channels" in this case hyperlinked to my instance results in a 500 error, not produced in the nginx logs and cannot find anything in the guincorn logs about it. stream is allowed with the stream key and is displayed as live but it gives an error on chrome and firefox "The media could not be loaded, either because the server or network failed or because the format is not supported."
This is including a mirrored copy of OSP on a separate server spun up to test
I've managed to get the channel pages working but users cannot register. By visiting the /register link and filling out the form it immediately results in a "Internal Server Error" with nothing in gunicorn, nginx or osp logs.
edit: I managed to get osp to log the error, here is a pastebin the user account is created but the "user" default role isn't applied to the account, it throws an error but keeps the user in the database without a role attached to it at all. you are not able to add additional roles to the created users.
I have run the database upgrade, fwiw
rolling back to the previous release fixes the issue on 7/7 attempts
when viewing a live stream that has not failed after server upgrade, vising "my channels" in this case hyperlinked to my instance results in a 500 error, not produced in the nginx logs and cannot find anything in the guincorn logs about it. stream is allowed with the stream key and is displayed as live but it gives an error on chrome and firefox "The media could not be loaded, either because the server or network failed or because the format is not supported."
This is including a mirrored copy of OSP on a separate server spun up to test
can be replicated by installing firefox on android and visiting the live channel, video says the playback format is not supported and apparently this is an open issue in regards to videojs as seen here
I'm not sure if there is a beta release of videojs or something that may rectify this issue?