Yes I had considered that, but considering that ours is not a giant but moderate instance (a thousand subscribers) it seemed exaggerated to get to have 1GB occupied every day.

well, indeed it could be just that. I just checked and it’s size is 15GB 😱

Thanks, will try it 👍

My problem was that the Hetzner server was new and therefore certain ports were closed. I solved it by contacting Hetzner support and having them open the ports.

It seems to have been resolved: the problem was the private instance setting, as soon as I edited in the database and it worked again 🙈

I actually only used Ansible for the first installation, all subsequent ones I manually updated only the docker file and following the various instructions that were given only in this update it didn’t work.

If it helps I can post my current configuration file docker-compose.yml

edit:

maybe I need to try again to do an update through Ansible now?

Since it had always worked to update the docker-compose.yml file it seemed easier and more functional to do it this way but I don’t want to think that maybe the installation is now “tied” to Ansible.

I had to restore an old backup because in no way had the problem resolved itself.

If you’re interested on GitHub they suggested this to me: https://github.com/LemmyNet/lemmy/issues/2653#issuecomment-1380541659 unfortunately the suggestion came after I restored the backup and I couldn’t try it.

If you’ll try it let me know if it works or not!

Hello, I have the exact same problem as you, and it also happened to me at the time when my instance was running out of free space.

Did you by any chance solve it somehow?

I tested from Mastodon but votes seems not works

Found out just now what the problem was.

I am writing this to help others: the federation was active in that instance but it was also marked as a private instance. This gave no problems in using it until the update (probably because it is restarted and re-run Docker).

I removed the federation from the config file leaving only the private instance and it restarted immediately.

I have answered and find the solution here: https://lemmy.ml/post/258218/comment/176759

Thanks for the help!

I think i got it solved.

It was fault of being a new customer of Hetzner, ask them to make some tests and with port opens it’ll go!

yes, actually i tried ‘none’, ‘tls’ and ‘starttls’ but with no success. I tried with external SMTP because the default created by ansible does not work:

  email: {
    smtp_server: "postfix:25"
    smtp_from_address: "noreply@mydomain.it"
    tls_type: "none"
  }

It is the same thing that I thought. Unfortunately, if I try to manually force the port by writing smtps.aruba.it:465 as I said the server freezes for several minutes, goes to 504 and does not send any email.

So I can not understand if the port must be entered in another way (?) or if I have to install something on the server first to be able to send an email through an external SMTP as in this case?

I was able to get it started/running with this command: ansible-playbook -vvv -u USER -k -K -i inventory/hosts lemmy.yml

Anyway now it ends with an error (it’s Debian 11):

if \“driver failed programming external connectivity\” in ex.explanation:\n", “invocation”: {“module_args”: {“project_src”: “/srv/lemmy/social.lealternative.net”, “state”: “present”, “pull”: true, “remove_orphans”: true, “docker_host”: “unix://var/run/docker.sock”, “tls_hostname”: “localhost”, “api_version”: “auto”, “timeout”: 10, “tls”: false, “validate_certs”: false, “debug”: false, “hostname_check”: false, “recreate”: “smart”, “build”: false, “remove_volumes”: false, “stopped”: false, “restarted”: false, “dependencies”: true, “nocache”: false, “ca_cert”: null, “client_cert”: null, “client_key”: null, “ssl_version”: null, “project_name”: null, “files”: null, “definition”: null, “remove_images”: null, “scale”: null, “services”: null}}}\n’, b"<stdin>:18: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module’s documentation for alternative uses\n") myuser@myip Failed to connect to the host via ssh: <stdin>:18: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module’s documentation for alternative uses The full traceback is: File “/tmp/ansible_docker_compose_payload_mgoivh5q/main.py”, line 777, in cmd_up self.project.up( File “/usr/lib/python3/dist-packages/compose/project.py”, line 565, in up results, errors = parallel.parallel_execute( File “/usr/lib/python3/dist-packages/compose/parallel.py”, line 112, in parallel_execute raise error_to_reraise File “/usr/lib/python3/dist-packages/compose/parallel.py”, line 210, in producer result = func(obj) File “/usr/lib/python3/dist-packages/compose/project.py”, line 548, in do return service.execute_convergence_plan( File “/usr/lib/python3/dist-packages/compose/service.py”, line 545, in execute_convergence_plan return self._execute_convergence_create( File “/usr/lib/python3/dist-packages/compose/service.py”, line 460, in _execute_convergence_create containers, errors = parallel_execute( File “/usr/lib/python3/dist-packages/compose/parallel.py”, line 112, in parallel_execute raise error_to_reraise File “/usr/lib/python3/dist-packages/compose/parallel.py”, line 210, in producer result = func(obj) File “/usr/lib/python3/dist-packages/compose/service.py”, line 465, in <lambda> lambda service_name: create_and_start(self, service_name.number), File “/usr/lib/python3/dist-packages/compose/service.py”, line 457, in create_and_start self.start_container(container) File “/usr/lib/python3/dist-packages/compose/service.py”, line 627, in start_container if “driver failed programming external connectivity” in ex.explanation:

fatal: [myuser@myip]: FAILED! => { “changed”: false, “errors”: [], “invocation”: { “module_args”: { “api_version”: “auto”, “build”: false, “ca_cert”: null, “client_cert”: null, “client_key”: null, “debug”: false, “definition”: null, “dependencies”: true, “docker_host”: “unix://var/run/docker.sock”, “files”: null, “hostname_check”: false, “nocache”: false, “project_name”: null, “project_src”: “/srv/lemmy/social.lealternative.net”, “pull”: true, “recreate”: “smart”, “remove_images”: null, “remove_orphans”: true, “remove_volumes”: false, “restarted”: false, “scale”: null, “services”: null, “ssl_version”: null, “state”: “present”, “stopped”: false, “timeout”: 10, “tls”: false, “tls_hostname”: “localhost”, “validate_certs”: false } }, “module_stderr”: “Creating sociallealternativenet_postgres_1 … \nCreating sociallealternativenet_postfix_1 … \nCreating sociallealternativenet_pictrs_1 … \n\u001b[3A\u001b[2K\nCreating sociallealternativenet_postgres_1 … \n\u001b[3B\u001b[1A\u001b[2K\nCreating sociallealternativenet_pictrs_1 … \ny_1 … \n\u001b[3A\u001b[2K\nCreating sociallealternativenet_postfix_1 … \n\u001b[3B”, “module_stdout”: “”, “msg”: “Error starting project a bytes-like object is required, not ‘str’” }

Be careful because Kiwi in the latest release insert a lot of tracking: https://reports.exodus-privacy.eu.org/it/reports/com.kiwibrowser.browser/latest/

• Facebook ADS
• Google AdMob
• Google Firebase

Same :) Try this instance: https://visionon.tv/w/3f29b19f-c0e7-4b0f-8bc2-2fab8da876f6

I understand doubts, especially if there is a lot of work to do I think it is right to ask as many questions as possible.

Personally I think of at least two other reasons for changing the instance:

the first is a possible problem with the moderators / admin. For example in case they did questionable operations.

The second concerns me more personally: I opened a community here because there are no Italian instances. It took me a few months to convince people to sign up here, it would be ironic to tell them “no, now re-sign up somewhere else”.

Since in this thread we talked about how to invite people to subscribe and / or how to create more instances for me this is one of the reasons that pushes me not to try.

In the meantime, thanks for the discussion and listening

People are often lazy, especially if I convinced them to join the “xxx instead of reddit” site to follow me. It would definitely be counterproductive to tell them again “ah no now you have to follow me on yyy, not xxx anymore”.

Obviously it is right to consider the work to be done and the right priorities. Personally I love Lemmy, I would be a little sorry to be somehow linked to this instance “forever”, I think the idea of ​​decentralization and faithful universe is a little less.

just my thoughts :)

Yes, on Mastodon there is an “export” and “import” settings for the followers (not for the post): https://docs.joinmastodon.org/user/moving/ and on the old istance remain just a placeholder that redirect on new profile/instances.

I’m glad the idea is of your interest :)