Add job for ticket garbage collection and implement stale ticket handling #4
Conversation
Summary of ChangesHello @tuhalf, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request establishes a robust system for managing the lifecycle of P2P ticket objects within the database. By introducing automatic garbage collection and defining clear criteria for ticket staleness, it ensures that the system remains efficient and free of outdated data, improving overall database health and performance. The changes encompass schema modifications, new background jobs, and integrated staleness checks during object retrieval. Highlights
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a valuable ticket expiration and garbage collection system for P2P objects, which will help maintain the database by cleaning up stale ticket objects. The implementation includes a database migration for a stored_at timestamp, a new cron job for periodic cleanup, and logic to filter stale objects on read. My review focuses on potential performance improvements and enhancing code clarity. I've identified that the prune_stale_objects function could lead to high memory usage with a large database and have suggested a more scalable batching approach. Additionally, I've provided suggestions to make some parts of the code more idiomatic and concise. Regarding your question about deleting tickets used for payments, the current logic correctly implements a cleanup strategy for a distributed cache. It assumes that tickets that are old (by time or epoch) are no longer relevant, which is a reasonable approach as submitted tickets are recorded on-chain.
|
Other than making pruning an sql command, fixed a bug on missing Kademlia objects. Handled missing Kademlia objects to stop decode! crashes: KBuckets.object now returns nil when absent, connection/RPC helpers skip missing objects, and edge/RPC responses filter out nil objects. |
dominicletz
left a comment
dominicletz
left a comment
There was a problem hiding this comment.
as per above we need to keep the network state consistent
lib/network/edge_v2.ex
Outdated
| ["getnodes", node] -> | ||
| KademliaLight.find_nodes(node) | ||
| |> Enum.map(&KBuckets.object/1) | ||
| |> Enum.filter(& &1) |
There was a problem hiding this comment.
nil filtering should not be needed
lib/network/rpc.ex
Outdated
| node: Json.prepare!(KBuckets.object(item), big_x: false), | ||
| retries: encode16(item.retries) | ||
| } | ||
| case KBuckets.object(item) do |
There was a problem hiding this comment.
I think if at all we should do nil checking insideKademliaLight.network() and if the in-memory table has any entries that are not in the db anymore we need to throw a fatal error and rebuild/clear the in-memory table
lib/kbuckets.ex
Outdated
| case Model.KademliaSql.object(key(item)) do | ||
| nil -> nil | ||
| binary -> Object.decode!(binary) | ||
| end |
There was a problem hiding this comment.
Same as below comment on KademliaLight.network() consistency -- if the db is not consistent this should crash and probably flush the memory table
lib/kademlia_light.ex
Outdated
| host = Server.host(server) | ||
| port = Server.peer_port(server) | ||
| Network.Server.ensure_node_connection(PeerHandlerV2, node_id, host, port) | ||
| case KBuckets.object(item) do |
lib/kademlia_light.ex
Outdated
| error -> | ||
| Logger.warning( | ||
| "Failed to get a result from #{Wallet.printable(node_id)} #{inspect(error)}" | ||
| case ensure_node_connection(node) do |
There was a problem hiding this comment.
Same as below, shouldn't be happening
|
I commited everything but nil check. I will complete it today |
lib/kademlia_light.ex
Outdated
|
|
||
| def network() do | ||
| call(fn _from, state -> {:reply, state.network, state} end) | ||
| call(&ensure_network_integrity/2) |
There was a problem hiding this comment.
network = call(fn _from, state -> {:reply, state.network, state} end)
if ensure_network_integrity(network) == false do
GenServer.cast(KademliaLight, :integrity_failed)
raise :network_integrity_failed
end
network
This pull request introduces a ticket expiration and garbage collection system for P2P objects, primarily focusing on cleaning up stale ticket objects from the database. It modifies the storage and retrieval logic for objects, adds new periodic cleanup jobs, and includes supporting tests to ensure stale tickets are properly expired and removed.
I am not sure about if I delete tickets that's used for payments or not. I need you to check that @dominicletz
Note
Introduce hourly GC that prunes stale ticket objects (tracked via new stored_at column) and drops corresponding Kademlia nodes, plus integrity checks and minor RPC listing tweak.
stored_attimestamp column top2p_objects(with migration) and store it on writes.prune_stale_objects/0to delete stale ticket rows in chunks and callKademliaLight.drop_nodes/1for their keys.network/0integrity check to remove nodes missing corresponding objects viadrop_nodes/1.:reset,:network, and{:drop_nodes, keys}; exposedrop_nodes/1and updatereset/0.Model.KademliaSql.prune_stale_objects/0.dio_networkto reverse bucket listing before merging with connected nodes.Written by Cursor Bugbot for commit 8c9a8f7. This will update automatically on new commits. Configure here.