Sessions transfer #5229
Sessions transfer #5229
Conversation
ruslandoga
force-pushed
the
sessions-transfer
branch
6 times, most recently
from
17a0173 to
96d1b27
Compare
There was a problem hiding this comment.
Copilot wasn't able to review any files in this pull request.
Files not reviewed (4)
- lib/plausible/application.ex: Language not supported
- lib/plausible/session/persistence.ex: Language not supported
- lib/plausible/session/persistence/tinysock.ex: Language not supported
- lib/plausible_web/controllers/api/system_controller.ex: Language not supported
ruslandoga
force-pushed
the
sessions-transfer
branch
8 times, most recently
from
6ddc0b7 to
e54aff6
Compare
ruslandoga
force-pushed
the
sessions-transfer
branch
from
8421233 to
f0ae022
Compare
|
|
| defp file_stat_ctime(path) do | ||
| case File.stat(path) do | ||
| {:ok, stat} -> stat.ctime | ||
| {:error, _} -> nil |
There was a problem hiding this comment.
What category of errors result with nil here? Shouldn't we then filter the path out and log failure?
There was a problem hiding this comment.
I haven't thought about it much, but the idea was to allow tinysock client to remove stale/broken files in sock_connect_or_rm and since tuple (normal stat.ctime) is > nil, this ordering is like "asc nulls last" so the cleanup happens in the end of the transfer.
| @@ -0,0 +1,31 @@ | |||
| defmodule Plausible.Session.Transfer.Alive do | |||
| @moduledoc false | |||
There was a problem hiding this comment.
It would be beneficial to keep the docs.
There was a problem hiding this comment.
I wanted to have only one place with the docs, Plausible.Session.Transfer.
It already briefly explains the role of this "alive" process.
Attempt at transferring in-memory sessions across deployments using shared persistent cache and Unix domain sockets.