Skip to main content

· 2 min read
John Lotoski

High level summary​

The SRE team continues work on Cardano environment improvements and general maintenance.

Some notable recent changes, updates or improvements include:

  • After upgrading hydra to new hardware and increasing eval concurrency for better ci throughput, an upstream hydra bug caused builds to frequently block in hydra's queue runner. The root caused was investigated and resolved with the fix to the issue pushed upstream.

  • A sizeable chunk of effort for this biweekly was invested in cardano-node's new tracing system, in particular between cardano-node, iohk-nix and cardano-parts repositories. Test machines were built and deployed using the latest tracing system changes intended for node 10.3 with comparisons between new tracing and legacy tracing metrics made and the few remaining discrepancies addressed by the performance and tracing team. New tracing system dashboards and alerts were prepared with recent changes. Configurations in each repo were prepared for a switch to the new tracing system as default. Readiness to migrate cardano-parts clusters to the new tracing system is close!

Repository Work -- Merged​

Cardano-node​

Repository Work In Progress -- PRs and Branches​

· One min read
Alexey Kuleshevich

High level summary​

Few minor touchups were implemented for the upcoming cardano-node-10.3 release. Other than that focus was mostly on removing redundant complexity and improvements to our test suites.

Low level summary​

Features​

  • pull-4924 - Make instantStake deserialization backwards compatible for Conway era
  • pull-4933 - Add required CBOR instances for NonZero
  • pull-4926 - Remove bytestring from Block type
  • pull-4938 - Remove redundant bytes memoization from WitVKey and BootstrapWitness

Testing​

  • pull-4944 - Implement more Alonzo UTxOW Imp tests
  • pull-4953 - Implement the remaining Alonzo UTxOW test and remove the old module

Infrastructure and releasing​

  • pull-4942 - Add bounds on cardano-crypto-wrapper in packages that also use crypton
  • pull-4946 - Post-release bump versions in changelogs and cabal-files
  • pull-4951 - Add a shellcheck workflow to GitHub CI

· 2 min read
Jean-Philippe Raynaud

High level overview​

This week, the Mithril team created the pre-release for the 2513.0-pre distribution, which provides support for Cardano node v.10.2.1 and includes various bug fixes and improvements.

They implemented the aggregator's prototype 'slave' mode to support signer registration across multiple aggregators and started adapting the infrastructure to support it. They also focused on signing ancillary files in the Cardano database snapshots with an IOG key and worked on recording the origin of requests made to the aggregator API.

Finally, the team enhanced the handling of some aggregator functional errors to improve the accuracy of the signer metrics and kept refactoring the cryptographic library to improve clarity and maintainability.

Low level overview​

  • Pre-released the new distribution 2513.0-pre
  • Completed the issue E2E tests adaptation for multiple aggregators #2361
  • Completed the issue Use a macro to implement the Source trait #2365
  • Completed the issue Upgrade to Cardano 10.2 #2333
  • Completed the issue Some error messages should be considered as warning in signer registration to aggregator #2363
  • Worked on the issue Sign ancillary files with IOG key #2362
  • Worked on the issue Release 2513 distribution #2332
  • Worked on the issue Record origin of client requests in metrics #2382
  • Worked on the issue Re-organize STM library structure #2369

· One min read
Noon van der Silk

High-level summary​

The team is very excited to have relased the work on reducing the memory footprint of the Hydra node by merging #1618. This has been tested and appears to make a huge impact to long-running Hydra nodes with many transactions. We invite the community to test it more! We also made a couple of hotfixes and documentation improvements.

What did the team achieve?​

  • Merged work on reduced memory footprint for running a Hydra Node #1618
  • Progress on "side-load" of a snapshot #1858
  • Updated documentation for persistence cleaning #1899
  • Fixed a bug where the TUI wasn't getting enough history #1901
  • Allowed for etcd arguments to be passed on #1891; useful for controlling etcd more explicitly.
  • Added /snapshot/last-seen endpoint to query the latest snapshot #1886

What's next?​

  • Finish side-loading snapshots #1858
  • Tighten security options of the networking layer #1867
  • Publishing scripts with blockfrost #1668