[Merged by Bors] - fetch/peers: make latency based selection biased towards last requests #5688
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
|
problem in the test peers stick for too long to a couple of peers, that peer drops them but they can't make progress and eventually go out of sync. unsynced node can't submit transaction, hence the failure
|
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## develop #5688 +/- ##
=========================================
- Coverage 79.8% 79.7% -0.2%
=========================================
Files 279 279
Lines 28426 28468 +42
=========================================
- Hits 22712 22704 -8
- Misses 4148 4199 +51
+ Partials 1566 1565 -1 ☔ View full report in Codecov by Sentry. |
|
bors try |
|
bors cancel |
|
bors try |
tryAlready running a review |
tryBuild failed: |
dshulyak
changed the title
|
bors try |
tryBuild failed: |
fasmat
approved these changes
Mar 12, 2024
Comment on lines
599
to
600
| go func() { | ||
| data, err := f.sendBatch(peer, batch) |
There was a problem hiding this comment.
Not part of this PR, but this part of the code seems like it can be improved:
- the spawned go routine isn't tracked via an
errgroup.Grouporsync.Waitgroup - there is no cancellation via a
context.Contextin place, the caller has no control over when sending a request shall be aborted (shutdown or timeout) handleHashErrorlocks a mutex for its whole lifetime, whilereceiveResponsedoes not. So if something is wrong and a lot of errors happen everything might slow down 🤔
|
bors try |
tryBuild failed: |
|
bors merge |
spacemesh-bors bot
pushed a commit
that referenced
this pull request
Mar 12, 2024
#5688) in the flaky system test it took 4 minutes to stop sending requests to a node that was actively dropping requests. in this pr i switched peer latency estimator to be biased towards latency observed in last requests, i looked up if/how lotus node handles similar issues. beside that there will be an INFO log with top peers stats, and global average latency in nanoseconds and total number of peers. by default node will emit log every 30m and it can be tuned by adding log-peer-stats-interval in fetch section of the config. ```json "fetch": { "log-peer-stats-interval": "1m" } ```
|
Build failed: |
|
bors merge |
spacemesh-bors bot
pushed a commit
that referenced
this pull request
Mar 12, 2024
#5688) in the flaky system test it took 4 minutes to stop sending requests to a node that was actively dropping requests. in this pr i switched peer latency estimator to be biased towards latency observed in last requests, i looked up if/how lotus node handles similar issues. beside that there will be an INFO log with top peers stats, and global average latency in nanoseconds and total number of peers. by default node will emit log every 30m and it can be tuned by adding log-peer-stats-interval in fetch section of the config. ```json "fetch": { "log-peer-stats-interval": "1m" } ```
|
Build failed (retrying...): |
spacemesh-bors bot
pushed a commit
that referenced
this pull request
Mar 12, 2024
#5688) in the flaky system test it took 4 minutes to stop sending requests to a node that was actively dropping requests. in this pr i switched peer latency estimator to be biased towards latency observed in last requests, i looked up if/how lotus node handles similar issues. beside that there will be an INFO log with top peers stats, and global average latency in nanoseconds and total number of peers. by default node will emit log every 30m and it can be tuned by adding log-peer-stats-interval in fetch section of the config. ```json "fetch": { "log-peer-stats-interval": "1m" } ```
|
Pull request successfully merged into develop. Build succeeded:
|
spacemesh-bors
bot
changed the title
4 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.
in the flaky system test it took 4 minutes to stop sending requests to a node that was actively dropping requests.
in this pr i switched peer latency estimator to be biased towards latency observed in last requests, i looked up if/how lotus node handles similar issues.
beside that there will be an INFO log with top peers stats, and global average latency in nanoseconds and total number of peers. by default node will emit log every 30m and it can be tuned by adding log-peer-stats-interval in fetch section of the config.