Bug Fix: Fixed stream switch metadata (url, user_agent, stream_id, m3u_profile) being written to Redis before the switch was confirmed to succeed

This commit is contained in:
SergeantPanda 2026-04-12 12:02:20 -05:00
parent 65d14644b4
commit b629836b3d
3 changed files with 40 additions and 19 deletions

View file

@ -49,7 +49,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
### Fixed
- Fixed manual stream selection from the Stats page not enforcing M3U profile connection limits in multi-worker deployments. When a non-owning worker handled the `change_stream` request it correctly packaged `stream_id` and `m3u_profile_id` into the Redis pubsub message, but the owning worker's pubsub handler only consumed `url` and `user_agent` silently dropping both IDs before calling `stream_manager.update_url()`. Because `update_url` only calls `update_stream_profile()` when a `stream_id` is provided, the `profile_connections` counter was never updated after the switch, causing subsequent capacity checks to see incorrect counts and bypass the full-profile guard. The handler now extracts `stream_id` and `m3u_profile_id` from the event and forwards them to `update_url()`.
- Fixed manual stream selection from the Stats page not enforcing M3U profile connection limits in multi-worker deployments. When a non-owning worker handled the `change_stream` request it correctly packaged `stream_id` and `m3u_profile_id` into the Redis pubsub message, but the owning worker's pubsub handler only consumed `url` and `user_agent` silently dropping both IDs before calling `stream_manager.update_url()`. Because `update_url` only calls `update_stream_profile()` when a `stream_id` is provided, the `profile_connections` counter was never updated after the switch, causing subsequent capacity checks to see incorrect counts and bypass the full-profile guard. The handler now extracts `stream_id` and `m3u_profile_id` from the event and forwards them to `update_url()`. The bug did not affect single-worker / dev-mode deployments because the owning worker handles those requests directly without pubsub.
- Fixed the `next_stream` rotation endpoint applying the same class of bug: `get_stream_info_for_switch()` was called and returned `m3u_profile_id`, but the result was dropped when forwarding to `ChannelService.change_stream_url()`, so `update_stream_profile()` was never called and `profile_connections` counters were not updated after an automatic stream rotation.
- Fixed stream switch metadata (`url`, `user_agent`, `stream_id`, `m3u_profile`) being written to Redis before the switch was confirmed to succeed. If the switch failed, URL unchanged or exception during teardown, Redis described a URL not actually in use. Metadata is now written only after `update_url()` returns `True`; on failure the owner writes `stream_manager.url` back as the ground truth. The non-owner no longer pre-writes metadata at all, all needed info is carried in the pubsub payload and written by the owner after confirmation.
- Fixed uploading a local M3U file with no expiration date set sending the string `"null"` as the `exp_date` field in the `FormData` request, causing a 400 validation error from the API. Null/undefined values are now skipped when building the `FormData` body, matching the behaviour already present in the update path.
- Fixed `PATCH /api/channels/channels/edit/bulk/` returning a 500 error when the request body included a `streams` list. The bulk edit handler was iterating `validated_data` directly and calling `setattr(channel, "streams", value)`, which Django prohibits on ManyToMany fields. Also added an `@extend_schema` decorator so the Swagger UI correctly documents the endpoint as accepting a JSON array and shows the `streams` field. (Fixes #883)
- Fixed several incorrect or incomplete OpenAPI (`@extend_schema`) schemas across the API:

View file

@ -231,14 +231,8 @@ class ProxyServer:
event_m3u_profile_id = data.get("m3u_profile_id")
if new_url and channel_id in self.stream_managers:
# Update metadata in Redis
# Mark the switch as in-progress in Redis so other workers know to wait
if self.redis_client:
metadata_key = RedisKeys.channel_metadata(channel_id)
self.redis_client.hset(metadata_key, "url", new_url)
if user_agent:
self.redis_client.hset(metadata_key, "user_agent", user_agent)
# Set switch status
status_key = RedisKeys.switch_status(channel_id)
self.redis_client.set(status_key, "switching")
@ -249,6 +243,13 @@ class ProxyServer:
if success:
logger.info(f"Stream switch initiated for channel {channel_id}")
# Confirm the URL in metadata now that the switch happened
if self.redis_client:
metadata_key = RedisKeys.channel_metadata(channel_id)
self.redis_client.hset(metadata_key, "url", new_url)
if user_agent:
self.redis_client.hset(metadata_key, "user_agent", user_agent)
# Publish confirmation
switch_result = {
"event": EventType.STREAM_SWITCHED, # Use constant instead of string
@ -268,6 +269,14 @@ class ProxyServer:
else:
logger.error(f"Failed to switch stream for channel {channel_id}")
# Roll back the URL in metadata to what the manager will
# actually reconnect to. The non-owner may have pre-written
# the desired URL; use stream_manager.url (the ground truth)
# so Redis is consistent with the live stream.
if self.redis_client:
metadata_key = RedisKeys.channel_metadata(channel_id)
self.redis_client.hset(metadata_key, "url", stream_manager.url)
# Publish failure
switch_result = {
"event": EventType.STREAM_SWITCHED,

View file

@ -168,15 +168,6 @@ class ChannelService:
else:
result = {'status': 'success'}
# Update metadata in Redis regardless of ownership
if proxy_server.redis_client:
try:
ChannelService._update_channel_metadata(channel_id, new_url, user_agent, stream_id, m3u_profile_id)
result['metadata_updated'] = True
except Exception as e:
logger.error(f"Error updating Redis metadata: {e}", exc_info=True)
result['metadata_updated'] = False
# If we're the owner, update directly
if proxy_server.am_i_owner(channel_id) and channel_id in proxy_server.stream_managers:
logger.info(f"This worker is the owner, changing stream URL for channel {channel_id}")
@ -187,14 +178,33 @@ class ChannelService:
success = manager.update_url(new_url, stream_id, m3u_profile_id)
logger.info(f"Stream URL changed from {old_url} to {new_url}, result: {success}")
# Update Redis metadata based on the actual outcome.
# On success, write the new values. On failure, restore whatever URL
# the manager will actually reconnect to (may be old_url if the
# exception happened before self.url was reassigned, or new_url if it
# happened after) so Redis never describes a URL that isn't in use.
if proxy_server.redis_client:
try:
if success:
ChannelService._update_channel_metadata(channel_id, new_url, user_agent, stream_id, m3u_profile_id)
else:
ChannelService._update_channel_metadata(channel_id, manager.url, user_agent)
result['metadata_updated'] = True
except Exception as e:
logger.error(f"Error updating Redis metadata: {e}", exc_info=True)
result['metadata_updated'] = False
result.update({
'direct_update': True,
'success': success,
'worker_id': proxy_server.worker_id
})
else:
# If we're not the owner, publish an event for the owner to pick up
logger.info(f"Not the owner, requesting URL change via Redis PubSub")
# Not the owner: publish the switch event. The owner will update metadata
# after the actual switch attempt succeeds (or roll back on failure).
# All needed info (url, user_agent, stream_id, m3u_profile_id) is carried
# in the pubsub message, so there is no reason to pre-write metadata here.
logger.debug(f"This worker is not the owner, publishing stream switch event for channel {channel_id}")
if proxy_server.redis_client:
ChannelService._publish_stream_switch_event(channel_id, new_url, user_agent, stream_id, m3u_profile_id)
result.update({