Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

If one server is down, no new Server can connect #155

Closed
fredo opened this issue Apr 20, 2020 · 1 comment · Fixed by #185
Closed

If one server is down, no new Server can connect #155

fredo opened this issue Apr 20, 2020 · 1 comment · Fixed by #185
Assignees

Comments

@fredo
Copy link
Contributor

fredo commented Apr 20, 2020

Description

Because of some unfortunate dependency chain a new Server will not run if one server is down. This can be avoided though.
The reason for this following:

  • The room ensurer is supposed to join the broadcast rooms from the first server in the list
  • It will also connect to all servers to ensure rooms. If one server is down the room ensurer will fail and not join the rooms
  • Neither PFS, MS nor Matrix will function without the fetched broadcast rooms

This can be easily avoided if we loosen the policy if a server is not available.

Suggestion

Since the room ensurer always ensures the roomwith the first server in the list, it is only necessary to connect to that server. The room ensurer can still ensure the rooms from the others but will not break if the server is not available. If the first server is offline it could use the second to ensure rooms, and so on.

@fredo
Copy link
Contributor Author

fredo commented Aug 10, 2020

I looked again into this issue and found the following improvements (regardless any improvement which comes with #180)

  • do not break if server cannot connect to a server which is down
  • add admin power levels for new servers
  • remove admin power levels for removed servers This is not possible. Same power levels cannot change each other
  • try to join broadcast rooms beginning at the first server in the list, if down then second, and so on...
  • RE crashes at first run due to setting power levels

The question comes up, do we need to ensure broadcast rooms of other servers than the first reachable one? If I recall correctly it only gives a warning and since we only ensure with the first server we do not act other than printing the warning.

@ulope feel free to add points here

@fredo fredo self-assigned this Aug 11, 2020
This was linked to pull requests Aug 17, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
1 participant