Replace opentelemetry-jaeger with opentelemetry-otlp as a more flexible alternative. #7563
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
matklad
reviewed
Sep 6, 2022
| ] | ||
|
|
||
| [[package]] | ||
| name = "prost-build" |
There was a problem hiding this comment.
Huh, it makes no sense I think that we pull code generation here... I would expect the rust code to be pre-generated...
Will ask tonic folks about this.
There was a problem hiding this comment.
|
I’m assuming that I’m not needed here any more. ;) |
nagisa
approved these changes
Sep 6, 2022
This consolidates three unsafe blocks into a single location.
This is a pure refactor, including two main refactors: - replace `TrieCacheFactory` with `TrieConfig` - move config constants into `TrieConfig` The motivation is to facilitate wider configuration and changes to the config format in follow-up PRs. See also #7564.
This PR replaces an unmaintained fork of secp256k1 to the latest version of the original library that already have all required features with a more clear API. Also this will solve compilation issues with llvm 13.* tools on MacOS like the next: rust-lang/rust-bindgen#1834
Firstly, change is_on_current_chain method to return boolean status rather than error if block is not on current chain. The method is used in two places: in one it’s only used as a boolean predicate so returned error is irrelevant; in the other one the caller performs another check and duplicates code for creating the error. Changing is_on_current_chain return bool simplifies both call sites. Secondly, rewrite check_block_final_and_canonical to accept a list of headers rather than operating on single block hash. (With that also change ‘block’ in name into ‘blocks’). This has a few motivations: i) it makes last final block header being fetched only once, ii) since caller needs access to the block header this makes it so that header is fetched only once and iii) moving fetching of the headers to the caller will actually help slightly in upcoming efforts for cold storage.
The database corruption issue when disk is running out of space has been fixed upstream in facebook/rocksdb#6316. There’s no reason for us to carry our own hacky workaround.
#7574) This does three relate things: 1. When opening the database in read-only mode, StoreOpener now checks whether database version is the one node is built for. If it isn’t, opening fails. This is because we cannot perform migrations in read-only mode so there’s nothing we can do with the database. 2. When creating the database, initialise it by setting its version to the one node was built for. This moves the code from nearcore. 3. Introduce StoreOpener::create which refuses to open already existing database. Motivation here is mostly to remove database existence check from recompress_storage and thus get rid of the check_if_exists method to shrink StoreOpener API.
This is a great help in debugging, and I don't think there's much downside to doing that? Capturing a backtrace is somewhat costly, especially on windows, but given that we run with panic=abort, that's bound to be a pretty cold path anyways! To disable backtraces, set `RUST_BACKTRACE=0` env variable. Test Plan --------- Manually check that both default&override work
Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com>
Now it is straightforward to see how messages are read from the network socket.
* added tests for connection pool * Update chain/network/src/broadcast/mod.rs Co-authored-by: Simonas Kazlauskas <git@kazlauskas.me> * applied comments * compilation fix Co-authored-by: Simonas Kazlauskas <git@kazlauskas.me> Co-authored-by: near-bulldozer[bot] <73298989+near-bulldozer[bot]@users.noreply.github.com>
cc: @matklad ``` % cargo bench -p near-o11y Compiling near-o11y v0.0.0 (/storage/code/nearcore-master/core/o11y) Finished bench [optimized] target(s) in 1.05s Running unittests src/lib.rs (target/release/deps/near_o11y-ccbb448e66a5a4d9) running 0 tests test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s Running benches/metrics.rs (target/release/deps/metrics-d531efabdee84e90) running 3 tests test inc_counter_vec_cached ... bench: 21 ns/iter (+/- 1) test inc_counter_vec_cached_str ... bench: 183 ns/iter (+/- 1) test inc_counter_vec_with_label_values ... bench: 506 ns/iter (+/- 10) test result: ok. 0 passed; 0 failed; 0 ignored; 3 measured ```
resolves #5182 nayduck run: https://nayduck.near.org/#/run/2664
…re (#7569) # account_id_in_function_call_permission This feature controls [the check](https://github.com/near/nearcore/blob/b315192e84d388671316deaa3a17ece9d0565fd1/runtime/runtime/src/verifier.rs#L400-L405) which enforces that account id in function call permission is indeed a valid account id. Before, any string could have been used there. The primary motivation is robustness -- by restricting permissions to only valid account ids, we don't have to deal with potentially arbitrary long strings in storage. # Context - Implementation: #7139 # Testing and QA We have basic and upgradability test [here](https://github.com/near/nearcore/blob/master/integration-tests/src/tests/client/features/account_id_in_function_call_permission.rs#L18). This PR also adds a test for an extra edge case with overly long account id. This feature have been running on betanet for couple of months without problems. # Checklist - [x] Link to nightly nayduck run: https://nayduck.near.org/#/run/2667 - [x] Update CHANGELOG.md to include this protocol feature in the `Unreleased` section.
Now it is straightforward to see how messages are read from the network socket.
nikurt
added a commit
that referenced
this pull request
Nov 9, 2022
…le alternative. (#7563) cc: @vlad-kopcil
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.
cc: @vlad-kopcil