Implement zeroconf for operational node discovery #531
Conversation
What polling behavior are you referring to? The rescheduled interview when a node stopped being available (not resolving) was on our end no? |
| # Remove and cancel any existing interview/subscription reschedule timer | ||
| if existing := self._sub_retry_timer.pop(node_id, None): | ||
| existing.cancel() | ||
| # shutdown existing subscriptions | ||
| if sub := self._subscriptions.pop(node_id, None): | ||
| await self._call_sdk(sub.Shutdown) |
There was a problem hiding this comment.
This code is also used when deleting the node, I think we should extract it in a common function.
| self.server.signal_event(EventType.NODE_UPDATED, node) | ||
| # NOTE: if the node is (re)discovered by mdns, that callback will | ||
| # take care of resubscribing to the node | ||
| asyncio.create_task(self._node_offline(node_id)) |
There was a problem hiding this comment.
I am a bit worried that we get false positives here. But let's see.
Btw, you should store the task somewhere, see also https://docs.python.org/3/library/asyncio-task.html#asyncio.create_task and https://bugs.python.org/issue44665.
Important: Save a reference to the result of this function, to avoid a task disappearing mid-execution. The event loop only keeps weak references to tasks. A task that isn’t referenced elsewhere may get garbage collected at any time, even before it’s done. For reliable “fire-and-forget” background tasks, gather them in a collection:
Implemented zeroconf to do our own mdns discovery of operational (and later commissionable) Matter nodes.
This solves a number of issues, most importantly that we do no longer have to rely on the (bad) polling behavior of the SDK if a node goes unavailable. Instead we can detect it coming back online through mdns discovery proactively.
Fixes: