Introduce NW BatchV2 in Protocol Version 12 & Pipe ProtocolConfig into NW #12178
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
arun-koshy
requested review from
akichidis,
bmwill,
asonnino and
lanvidr
as code owners
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
mwtian
reviewed
May 24, 2023
mwtian
approved these changes
May 24, 2023
| // TODO: Remove once we have upgraded to protocol version 12. | ||
| if self.protocol_config.narwhal_versioned_metadata() { | ||
| // Set received_at timestamp for remote batches. | ||
| let mut updated_new_batches = HashMap::new(); |
There was a problem hiding this comment.
Is it possible to make new_batches mut and update them inplace? It would avoid the copy and maybe a bit of duplicated logic.
There was a problem hiding this comment.
I had issues trying to do this before which is why I had to do it this way, but I will follow up in a separate PR if I can get it to work in place
akichidis
approved these changes
May 24, 2023
|
Reply |
arun-koshy
force-pushed
the
ak/nw-versioning
branch
from
4c442e1
to
f201df1
Compare
arun-koshy
force-pushed
the
ak/nw-versioning
branch
from
dfc871c
to
f8d79f6
Compare
arun-koshy
force-pushed
the
ak/nw-versioning
branch
from
f8d79f6
to
10037d3
Compare
mwtian
approved these changes
May 27, 2023
…NW VersionedMetadata field
and add epoch to BatchV2 to be included in digest
Fix tests
Remove epoch from BatchV2 Fix rustfmt errors
arun-koshy
force-pushed
the
ak/nw-versioning
branch
from
49696b7
to
3829c6a
Compare
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Outdated suggestions cannot be applied.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.
Description
This is attempt 2 of getting protocol config into narwhal. Previous attempt (PR#11519) had to be reverted because mismatched batch versions were causing the validators to panic. The issue was that we were not handling the protocol upgrade correctly.
ProtocolConfigwas passed in whenNarwhalManagerwas created but on epoch changemonitor_reconfigurationdoes not recreateNarwhalManagerif it still has a handle on theValidatorComponentsbut rather just callsstartfrom the existingNarwhalManager. This is not a problem forNarwhalConfigurationparameters which is also only passed in onNarwhalManagercreation because the moment the node is restarted for a binary update the parameters take effect. However in the case of protocol upgradesProtocolConfigis only updated on the following epoch.For example if the validator restarts to update its binary from version
Nto versionN+1,NarwhalManagerwould be constructed withProtocolConfigat versionN(notN+1) because we need to ensure we have a majority quorum before actually going to versionN+1which happens at epoch change. On epoch change because the node still has a handle onValidatorComponentsit just starts Narwhal with the existingNarwhalManagerand ends up usingProtocolConfigat versionNwhich is the root cause of the issues.To resolve this the following changes were added to the last PR to fix the issue and make it more robust
ProtocolConfigtoNarwhalManageron start and not on creationTest Plan
Added unit tests & tested protocol upgrade in labnet from mainnet release branch to v12
Type of Change (Check all that apply)
Release notes
Start using
BatchV2in Narwhal which introducesVersionedMetadatathat allows for more granular tracking of NW batch execution latency.