Releases: stellar/stellar-core
Releases · stellar/stellar-core
v18.5.0
Stability improvements
- closeLedger: only call sizeOp once
- Improve overlay logging
- Bump overlay min version to 19
- Stop deleting dead entries for older buckets
- Update the
quorumendpoint logic - Fix 'BucketListIsConsistentWithDatabase added entries'
- Report flow control info in "peers" endpoint
- Improve overlay idle timeout
- Make dropping sqlite databases fast
- Configure error if ccache enabled but not found
- Add brew install ccache to macos instructions
- Improve instructions for installing on macOS
- Clean up inconsistent usage of config::CURRENT_LEDGER_PROTOCOL_VERSION
New features
- Improve config defaults and other minor changes Enable overlay flow control by default
Breaking changes
- Update project files to Visual Studio 2022
- Stricter offline info; With this change, stellar-core sends logs to stderr instead of stdout
v18.5.0rc1
Stability improvements
- closeLedger: only call sizeOp once
- Improve overlay logging
- Bump overlay min version to 19
- Stop deleting dead entries for older buckets
- Update the
quorumendpoint logic - Fix 'BucketListIsConsistentWithDatabase added entries'
- Report flow control info in "peers" endpoint
- Improve overlay idle timeout
- Make dropping sqlite databases fast
- Configure error if ccache enabled but not found
- Add brew install ccache to macos instructions
- Improve instructions for installing on macOS
- Clean up inconsistent usage of config::CURRENT_LEDGER_PROTOCOL_VERSION
New features
- Improve config defaults and other minor changes Enable overlay flow control by default
Breaking changes
- Update project files to Visual Studio 2022
- Stricter offline info; With this change, stellar-core sends logs to stderr instead of stdout
v18.4.0
Release notes:
Stability improvements
- Overlay flow control not enabled by default for now as it’s a major update to how peers communicate with each other for better performance regardless of peer types.
- Clear state if stellar-core aborts while applying buckets this fixes an edge case that could lead the ledger in a bad state if the node was stopped at the wrong time.
- Log as much uncaught-exception info as we can, fixes #3155.
- Actually enforce SHA256 contract
- Rename maxTxSize to maxTxSetSize.
- Cleanup logic for checking protocol versions.
- Rename isString32Valid to isStringValid
- Log exception as INFO (instead ERROR) for older protocol version
- Only git clean ignored files in Dockerfile.testing
- Simulation updates and fixes
- Add TrustFlagsOpFrameBase to Visual Studio project.
- Return early in updateEntryIfRecorded if entry is not an offer
- Update catch2 to 2.13.8
- Unify trustline op
- Bump spdlog1.9.2
New features
- Make validators from the same home domain preferred peers. This implements a best practice that validators within the same organization should utilize internal bandwidth when possible.
- Adding an option to halt on txINTERNAL_ERROR. - this allows to configure certain nodes to never externalize ledgers that may have run into unexpected problems.
Breaking changes
v18.4.0rc1
Release notes:
Stability improvements
- Overlay flow control not enabled by default for now as it’s a major update to how peers communicate with each other for better performance regardless of peer types.
- Clear state if stellar-core aborts while applying buckets this fixes an edge case that could lead the ledger in a bad state if the node was stopped at the wrong time.
- Log as much uncaught-exception info as we can, fixes #3155.
- Actually enforce SHA256 contract
- Rename maxTxSize to maxTxSetSize.
- Cleanup logic for checking protocol versions.
- Rename isString32Valid to isStringValid
- Log exception as INFO (instead ERROR) for older protocol version
- Only git clean ignored files in Dockerfile.testing
- Simulation updates and fixes
- Add TrustFlagsOpFrameBase to Visual Studio project.
- Return early in updateEntryIfRecorded if entry is not an offer
- Update catch2 to 2.13.8
- Unify trustline op
- Bump spdlog1.9.2
New features
- Make validators from the same home domain preferred peers. This implements a best practice that validators within the same organization should utilize internal bandwidth when possible.
- Adding an option to halt on txINTERNAL_ERROR. - this allows to configure certain nodes to never externalize ledgers that may have run into unexpected problems.
Breaking changes
v18.3.0
Stability improvements
- Captive core buffers ledgers on startup ; minimizes potential downtime when restarting captive-core
- Delete temporary files while catching up
- bump medida to pick up latest KMS fixes ; changed the way percentiles are calculated to be more accurate for both low and high frequency sources, this change also reduces the window used for percentile reporting from last 5 minutes to latest full 30 seconds window.
- Track Init vs Live in LedgerTxn ; general performance improvement
- Add additional offers index ; general performance improvement
- Shutdown fixes ; resolves potential issues during shutdown
- Fix typo BATCH_SIZE -> BATCHSIZE in CONTRIBUTING.md
- Use HashOfHash in LedgerHashUtils and InternalLedgerEntry - refactor
- Fix transaction flooding test
- update install instructions to reference Postgresql
- AMM fuzzer updates
- Document simulation configurations in
stellar-core_example.cfg
New features
- bump medida to pick up latest KMS fixes ; this change also added a new quantile
q=1.0that tracks the maximum value over the last 30 seconds time window.
Breaking changes
None
v18.3.0rc2
Stability improvements
- Captive core buffers ledgers on startup ; minimizes potential downtime when restarting captive-core
- Delete temporary files while catching up
- bump medida to pick up latest KMS fixes ; changed the way percentiles are calculated to be more accurate for both low and high frequency sources, this change also reduces the window used for percentile reporting from last 5 minutes to latest full 30 seconds window.
- Track Init vs Live in LedgerTxn ; general performance improvement
- Add additional offers index ; general performance improvement
- Shutdown fixes ; resolves potential issues during shutdown
- Fix typo BATCH_SIZE -> BATCHSIZE in CONTRIBUTING.md
- Use HashOfHash in LedgerHashUtils and InternalLedgerEntry - refactor
- Fix transaction flooding test
- update install instructions to reference Postgresql
- AMM fuzzer updates
- Document simulation configurations in
stellar-core_example.cfg
New features
- bump medida to pick up latest KMS fixes ; this change also added a new quantile
q=1.0that tracks the maximum value over the last 30 seconds time window.
Breaking changes
None
v18.3.0rc1
Stability improvements
- Captive core buffers ledgers on startup ; minimizes potential downtime when restarting captive-core
- Delete temporary files while catching up
- bump medida to pick up latest KMS fixes ; changed the way percentiles are calculated to be more accurate for both low and high frequency sources, this change also reduces the window used for percentile reporting from last 5 minutes to latest full 30 seconds window.
- Track Init vs Live in LedgerTxn ; general performance improvement
- Add additional offers index ; general performance improvement
- Shutdown fixes ; resolves potential issues during shutdown
- Fix typo BATCH_SIZE -> BATCHSIZE in CONTRIBUTING.md
- Use HashOfHash in LedgerHashUtils and InternalLedgerEntry - refactor
- Fix transaction flooding test
- update install instructions to reference Postgresql
- AMM fuzzer updates
- Document simulation configurations in
stellar-core_example.cfg
New features
- bump medida to pick up latest KMS fixes ; this change also added a new quantile
q=1.0that tracks the maximum value over the last 30 seconds time window.
Breaking changes
None
v18.2.0
This release is a stability release mostly aimed at improving performance of captive core and network overall.
Stability improvements
- Populate mEntryCache, performance improvement
- Prevent
mSyncingLedgersfrom growing indefinitely when out of sync, lowering memory consumptions while catching up. - Increment iterator before break, that would sometimes update database entries multiple time while applying buckets.
- Clean up MultiOrderBook when offer is removed - improving memory usage a little bit.
- Query optimization for InMemoryLedgerTxn, restoring proper performance for captive core instances.
- Update the logic in
LogSlowExecutionto correctly count dropped messages - Check test tx meta in GitHub CI. Improve detection of protocol level bugs in tests.
- Move to
FMT_STRINGin production code, catches format errors at compile time. - Fix libc++ versions in Dockerfile.testing, fixing link issues in container.
- Fix passive flag xdr comment.
- add a few updates to docs related to meta ingestion with captive-core
- properly convert some enum values when printing json, could crash in some situations
- Check for DB file existence in captive core fixing a potential race condition on startup.
- [Enable docker-from-docker inside devcontainer]
(#3267)
New features
- Dampen arb flooding, by favoring transactions that are both higher fee and submitted first (only when identified as “arbitrage traffic”, other transactions are still flooded exclusively based on fee).
Breaking changes
None
v18.2.0rc2
This release is a stability release mostly aimed at improving performance of captive core and network overall.
Stability improvements
- Populate mEntryCache, performance improvement
- Prevent
mSyncingLedgersfrom growing indefinitely when out of sync, lowering memory consumptions while catching up. - Increment iterator before break, that would sometimes update database entries multiple time while applying buckets.
- Clean up MultiOrderBook when offer is removed - improving memory usage a little bit.
- Query optimization for InMemoryLedgerTxn, restoring proper performance for captive core instances.
- Update the logic in
LogSlowExecutionto correctly count dropped messages - Check test tx meta in GitHub CI. Improve detection of protocol level bugs in tests.
- Move to
FMT_STRINGin production code, catches format errors at compile time. - Fix libc++ versions in Dockerfile.testing, fixing link issues in container.
- Fix passive flag xdr comment.
- add a few updates to docs related to meta ingestion with captive-core
- properly convert some enum values when printing json, could crash in some situations
- Check for DB file existence in captive core fixing a potential race condition on startup.
- [Enable docker-from-docker inside devcontainer]
(#3267)
New features
- Dampen arb flooding, by favoring transactions that are both higher fee and submitted first (only when identified as “arbitrage traffic”, other transactions are still flooded exclusively based on fee).
Breaking changes
None
v18.2.0rc1
This release is a stability release mostly aimed at improving performance of captive core and network overall.
Stability improvements
- Populate mEntryCache, performance improvement
- Prevent
mSyncingLedgersfrom growing indefinitely when out of sync, lowering memory consumptions while catching up. - Increment iterator before break, that would sometimes update database entries multiple time while applying buckets.
- Clean up MultiOrderBook when offer is removed - improving memory usage a little bit.
- Query optimization for InMemoryLedgerTxn, restoring proper performance for captive core instances.
- Update the logic in
LogSlowExecutionto correctly count dropped messages - Check test tx meta in GitHub CI. Improve detection of protocol level bugs in tests.
- Move to
FMT_STRINGin production code, catches format errors at compile time. - Fix libc++ versions in Dockerfile.testing, fixing link issues in container.
- Fix passive flag xdr comment.
- add a few updates to docs related to meta ingestion with captive-core
- properly convert some enum values when printing json, could crash in some situations
- Check for DB file existence in captive core fixing a potential race condition on startup.
- [Enable docker-from-docker inside devcontainer]
(#3267)
New features
- Dampen arb flooding, by favoring transactions that are both higher fee and submitted first (only when identified as “arbitrage traffic”, other transactions are still flooded exclusively based on fee).
Breaking changes
None