fix(resharding): storage costs #12661
Merged
+138
−11
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
Longarithm
force-pushed
the
rv3-storage-fix
branch
from
e20f38a to
1a9dcd6
Compare
wacban
approved these changes
Dec 23, 2024
core/store/src/trie/mod.rs
Outdated
| @@ -698,7 +698,7 @@ impl Trie { | |||
| storage, | |||
| memtries, | |||
| root, | |||
| charge_gas_for_trie_node_access: flat_storage_chunk_view.is_none(), | |||
| charge_gas_for_trie_node_access: false, | |||
There was a problem hiding this comment.
Technically speaking this should be something more like the following. This one breaks replayability but I don't really mind to be honest given we actually only replay using tools and not this code path.
!ProtocolVersion::FlatStorage.enabled(protocol_version).
| @@ -754,6 +841,15 @@ fn test_resharding_v3_base(params: TestReshardingParameters) { | |||
| } | |||
| trie_sanity_check.assert_state_sanity(&clients, expected_num_shards); | |||
| latest_block_height.set(tip.height); | |||
| let shard_layout = | |||
| client.epoch_manager.get_epoch_config(&tip.epoch_id).unwrap().shard_layout; | |||
| println!( | |||
There was a problem hiding this comment.
nit: tracing::info and use the structured logging by setting fields rather than putting them in string
tracing::info!(
last_block_hash=?tip.last_block_hash,
...
"block"
);
There was a problem hiding this comment.
nevermind, the testloop tests use println quite often
|
@shreyan-gupta Can you have a quick look? I don't want to be merging PR that I myself approved ;) |
shreyan-gupta
approved these changes
Jan 7, 2025
github-merge-queue
bot
removed this pull request from the merge queue due to failed status checks
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.
Fixing old storage costs inaccuracy, which caused a failure during some forknet experiment https://near.zulipchat.com/#narrow/channel/407288-core.2Fresharding/topic/forknet/near/489699725
The fix is one-liner,
charge_gas_for_trie_node_access: false,. Originally it wasn't the case because we had a protocol upgrade from trie to flat storage read costs, so we needed to compute costs differently. But since flat storage costs were enabled, and we started to use only the flat storage read costs. Moreover, later it became controlled by runtimeParameter::FlatStorageReads, but the original condition stayed.And now, when we do flat storage resharding, flat storage indeed doesn't exist for a while, which triggered trie costs for some blocks during resharding again. However, on chunk validation, costs were correct, so chain couldn't validate any chunks since resharding start.
I test this by calling a contract which reads a key and then writes key-value pair back. For the old code, key read charges more cost than it should, which causes InvalidOutcomesProof error.