Merge branch 'fix/dup-and-shutdown' into fix/fable-mediums
This commit is contained in:
commit
5af20f5891
1 changed files with 657 additions and 29 deletions
11
README.md
11
README.md
|
|
@ -190,8 +190,15 @@ It answers `451` only when the database itself fails.
|
|||
- Hard bounces count for 30 days: when the last one is older, the next one starts again from 1. A disabled member
|
||||
stays disabled until it is added again.
|
||||
- A relay that accepted the message is never asked again because `QUIT` went wrong.
|
||||
- Delivery is at-least-once: a crash between the relay accepting a message and the queue entry being
|
||||
removed can send one message to one recipient twice.
|
||||
- The same post is queued once. If the `250` for a post never reaches Postfix, Postfix delivers the post again; a list
|
||||
remembers each post it queued for an hour (by `Message-ID`, or by a hash of the whole message when it has none),
|
||||
answers `250` to a copy and does not queue it again (the log says so). An approved held post is always delivered and is
|
||||
remembered too, so a copy that arrives later is not.
|
||||
- On stop, the daemon refuses new connections and lets the LMTP sessions and API or web requests in flight finish, so a
|
||||
post being received gets its `250`. After 30 seconds it cuts whatever is still open.
|
||||
- What remains is at-least-once delivery across the relay: a crash between the relay accepting a message and the queue
|
||||
entry being removed can send one message to one recipient twice. A copy of a post that arrives more than an hour later
|
||||
is also delivered again.
|
||||
- `GET /api/v1/queue` and the Queue page show what is pending or failing, and a stuck delivery can be discarded.
|
||||
|
||||
## Admin UI
|
||||
|
|
|
|||
Loading…
Reference in a new issue