feat: Autosharding API for relay subscriptions #1983
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
You can find the image built from this PR at Built from 86fc087 |
|
You can find the experimental image built from this PR at |
SionoiS
requested review from
vpavlin,
jm-clius,
NagyZoltanPeter and
Ivansete-status
jm-clius
reviewed
Sep 6, 2023
There was a problem hiding this comment.
In general I think:
- subscribing, unsubscribing and publishing on the underlying gossipsub object should only happen in one place. We currently have multiple paths towards each of these functions, which will be very error-prone and requires multiple logic duplications. If we add another API call for e.g.
subscribe, it should simply do the necessary translations/conversions (e.g. determining the pubsub topic via autosharding) to call the "base"subscribe. - I think we should avoid mixed configuration here: people should either subscribe via content topics or directly on pubsub topic. In other words, once you've unsubscribed from all content topics matching a specific shard, that shard subscription can be safely removed without considering if we have configured a "direct" subscription too.
- this would make the auto-subscription methods a very light layer on top of the existing subscribe/unsubscribe functions and hopefully minimise the state we have to keep
|
It now only use |
Ivansete-status
approved these changes
Sep 11, 2023
jm-clius
approved these changes
Sep 11, 2023
SionoiS
added
E:1.2: Autosharding for autoscaling
# This is the 1st commit message: feat: Autosharding API for relay subscriptions # This is the commit message #2: Fix
Ivansete-status
added a commit
that referenced
this pull request
Sep 26, 2023
…erly (#2082) * libwaku.nim: unsubscribe -> unsubscribeAll to make it build properly We introduced a change in the next PR #1983 that made the `libwaku` to stop building properly (make libwaku) because a signature change in the `unsubscribe` proc of waku/waku_relay/protocol.nim. This commit is a temporary workaround to make it work, and not block any ongoing PR tests, although not exactly as it is expected because we are unsubscribing from all topics. * ci.yml: adds library/** to force tests when changes under that folder
12 tasks
10 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Added new API that take only content topics and use autosharding to subscribe to the correct shard.
You have to manage the relay handlers otherwise it would happen that unsubscribing from a shard would remove all content topic handlers.
You also have to not duplicate or remove the default handlers otherwise Filter and/or Archive would receive no or duplicate messages.
I don't like adding more code to
waku_nodemaybe adding a new component/sub-system would be better?Changes
waku_nodetrack subscriptions to avoid unsubbing from shard used by other content topics.waku_relaynow has the same proc than GossipSubunsubscribeAllwhich take no handler andunsubscribethat remove only the handler provided.Tracking #1936