feat(ICRC_Rosetta): DEFI-2452: Spawn blocking for SQLite operations #7695
+1,096
−841
Conversation
This file contains hidden or 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
…setta-sqlite-cache-size-limit-and-flushing # Conflicts: # rs/rosetta-api/local/cluster/README.md # rs/rosetta-api/local/cluster/deploy.sh # rs/rosetta-api/local/cluster/templates/rosetta-icrc.yaml # rs/rosetta-api/local/cluster/values.yaml
…cache-size-limit-and-flushing' into mathias-DEFI-2452-add-configurable-icrc-rosetta-batch-size # Conflicts: # rs/rosetta-api/icrc1/src/common/storage/storage_client.rs # rs/rosetta-api/icrc1/src/common/storage/storage_operations.rs # rs/rosetta-api/icrc1/src/config.rs # rs/rosetta-api/icrc1/src/main.rs # rs/rosetta-api/local/cluster/README.md # rs/rosetta-api/local/cluster/deploy.sh # rs/rosetta-api/local/cluster/templates/rosetta-icrc.yaml # rs/rosetta-api/local/cluster/values.yaml
…ize' into mathias-DEFI-2452-spawn-blocking-for-sqlite # Conflicts: # rs/rosetta-api/icrc1/src/common/storage/storage_client.rs
…setta-sqlite-cache-size-limit-and-flushing
…cache-size-limit-and-flushing' into mathias-DEFI-2452-add-configurable-icrc-rosetta-batch-size # Conflicts: # rs/rosetta-api/local/cluster/deploy.sh # rs/rosetta-api/local/cluster/templates/rosetta-icrc.yaml
…ize' into mathias-DEFI-2452-spawn-blocking-for-sqlite
Base automatically changed from
mathias-DEFI-2452-add-configurable-icrc-rosetta-batch-size
to
master
November 25, 2025 16:28
# Conflicts: # rs/rosetta-api/icrc1/src/common/storage/storage_client.rs # rs/rosetta-api/icrc1/src/config.rs # rs/rosetta-api/local/cluster/README.md # rs/rosetta-api/local/cluster/values.yaml
|
|
||
| // As long as there are blocks to be fetched, keep on iterating over the blocks in the database with the given BATCH_SIZE interval | ||
| while !rosetta_blocks.is_empty() { | ||
| let mut account_balances_cache: HashMap<Account, BTreeMap<u64, Nat>> = HashMap::new(); |
Contributor
Author
There was a problem hiding this comment.
Good catch, thanks! I had moved it in an earlier commit of the PR this PR was stacked on top of. I later reverted the change here since I realized it didn't make sense, but I must have missed it when merging in the latest changes from master and fixing merge conflicts. I moved it back outside the while loop now.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Change the synchronous functions in ICRC Rosetta that call SQLite to perform database I/O to be asynchronous and use
tokio::task::spawn_blockingto run the task on a thread in a thread pool, rather than using one of the few asynchronous worker threads. For ICRC Rosetta instances with many (tens of) tokens, this can help to avoid situations where the worker threads are all busy servicing long-running SQLite requests, and there are no available threads to serve e.g., simple endpoint requests such asnetwork/status, or for running the watchdog threads to see if some of the block synchronization threads have gotten stuck. Using a thread pool for spawning threads to serve the SQLite requests will use more memory, which needs to be taken into account when dimensioning the system where Rosetta is running. The features in the previously proposed PRs for limiting the SQLite cache size, and for reducing the balance sync batch size, can be used as a counter-balance.