v1.24.0
Overview
We are excited to introduce Curio v1.24.0, which includes several improvements, bug fixes, and new features to enhance the overall performance and usability of Curio. This release prepares Curio for upcoming Filecoin Network upgrade to NV24.
📖 You can read more about Curio in our documentation.
Compatibility
This release is compatible with Lotus versions v1.30.0-rc2 and Boost v2.4.0-rc1. For the latest compatibility details, refer to the Curio version compatibility matrix.
⭐ Highlights
☀️ New Features
- Introduced a retry wait mechanism for task retries, enhancing system stability and reducing the likelihood of repeated immediate failures.
- Added alerts to notify users when the chain synchronization is lagging, ensuring timely awareness of synchronization issues.
- The UI now shows node uptime, providing better insights into node performance and stability.
- Users can now reset the sectors state to before SDR from the UI. This allows resealing sectors from SDR if required at any stage before PreCommit message is sent.
- Updated Curio to be compatible with the upcoming NV24 network upgrade, ensuring seamless operation post-upgrade.
🔧 Enhanced Snap Deals Handling
- Added a button in the web UI to resume all stuck sectors in Snap Deals, improving operational efficiency.
- Fixed issues where Snap Deals did not correctly process non-Filplus deals, ensuring all deals are handled properly.
- Adjusted scheduling logic to only schedule tasks when necessary, reducing unnecessary resource usage during Snap Deals.
⚡ Faster Performance
- Improved task scheduling to initiate processes only when needed, enhancing overall system performance.
- Introduced an experimental feature for GPU overprovisioning to potentially enhance computational tasks.
- Modified the loading of native dependencies to with certain build flags to allow using Curio is a library.
🖥️ User Interface Updates
- Users can now copy the block CID directly from the winning stats table.
- Enhanced the PoRep pipeline UI to display when tasks are waiting, providing better visibility into processing states.
🏗️ Improvements
- Added more detailed resource logs to assist users in monitoring and troubleshooting system resource usage.
- Enhanced the multinode proxy for better performance and reliability in multi-node environments.
- Implemented transactions during sector migrations to ensure data integrity and consistency.
- Improved the sector termination process for enhanced reliability and user experience.
- Cleaned up code by removing unused wdPost pre-checks and configurations, reducing potential confusion.
🐛 Bug Fixes
- Fixed an issue where counters were not correctly separated when sharing max values, improving resource allocation accuracy.
- Eliminated incomplete resource check logs that could cause confusion or misinterpretation.
- Corrected the PoRep count in the summary to display accurate information.
- Fixed an issue where multiple API keys were found in the database when migrating multiple miner IDs.
- Fixed the sector list power display after sector extension operations to reflect accurate values.
- Fixed a bug where the message watcher could return early, ensuring consistent message processing.
- Addressed an issue causing panics due to faults by creating a skipped bitfield in PoSt operations.
- Resolved issues where Slack alerts and winpost alerts were not triggering correctly.
- Corrected alert mechanisms related to WindowPoSt and WinningPoSt operations.
What's Changed
- feat: implement retry wait mechanism for task retries by @strahe in #216
- feat: add github action by @LexLuthr in #229
- fix: harmony: Correctly separate this counters when sharing Max by @magik6k in #246
- fix: add detailed resource log by @LexLuthr in #205
- feat: node uptime in UI by @LexLuthr in #253
- fix: remove incomplete resource check log by @magik6k in #261
- feat: restart button for sectors by @LexLuthr in #266
- fix: PoRep count in PoRep summary by @LexLuthr in #257
- fix: show task waiting in PoRep pipeline by @LexLuthr in #260
- fix: multiple api keys found in the database when migrating multiple miner IDs by @LexLuthr in #256
- fix: allow copying the block cid from winning stats table by @LexLuthr in #255
- fix: remove wdPost pre-checks, unused config by @LexLuthr in #252
- fix: alerts for wnPost and wdPost by @LexLuthr in #251
- update sector terminate by @LexLuthr in #264
- remove dependabot by @LexLuthr in #274
- fix: sector list power after extension by @LexLuthr in #273
- feat: chain sync alert by @LexLuthr in #272
- fix: do not return early from message watcher by @LexLuthr in #276
- fix: post: create skipped bitfield to avoid panics with faults by @magik6k in #279
- fix: snap: Correctly handle non-filplus deals by @magik6k in #281
- fix: sealsupra: Only schedule when needed by @magik6k in #284
- feat: webui/snap: Button for resuming all stuck sectors by @magik6k in #230
- cunative: Only pull in native deps with a tag by @magik6k in #211
- experiment: GPU Overprovisioning by @magik6k in #262
- fix make calibnet by @magik6k in #285
- fix: slack alert, winpost alert by @LexLuthr in #287
- fix: Improved multinode proxy by @LexLuthr in #249
- Use transactions in sector migrations by @magik6k in #169
- feat: NV24 upgrade by @LexLuthr in #286
- chore: release v1.24.0 by @LexLuthr in #291
Full Changelog: v1.23.1...v1.24.0