Skip to main content

· 2 min read
John Lotoski

High level summary​

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

Some notable recent updates or improvements include:

  • Work on cardano-parts repo continues and was utilized by the sanchonet-demo repository for demonstration purposes
  • Work on a new capkgs repository started -- "Content Addressed Packages"
    • This repository provides common Cardano release and pre-release binaries in a pure content addressed nix package type via flake output
    • These packages don't have the associated IFD, long eval times and large source dependencies that the originating source repositories often do
    • Consumers of such packages, such as devShells or nixos machines, can benefit from the improved performance this offers

Lower level summary​

Capkgs​

  • Content addressed packages -- new repo: capkgs

Cardano-ops​

  • Cardano p2p relay updates, dbSync snapshots server update, monitoring adjustments: cardano-ops-compare

Cardano-parts​

Cardano-perf​

Cardano-world​

Sanchonet-demo​

· 2 min read
Carlos LopezDeLara

2023-08-16 - 2023-08-29

High level summary​

CLI is making progress towards phases 3 and 4 of SanchoNet.

  • Ability to create info, no-confidence and new committee governance actions.

  • Voting as DREP is possible now on the CLI

  • API is now using Ledger's PParams and VotingProcedures

  • Expose following queries from consensus:

    - GetGovState
    - GetDRepState
    - GetDRepStakeDistr
    - GetCommitteeState
    - GetConstitution

cardano-cli​

cardano-api​

cardano-node​

cardano-testnet​

docs​

CI & project maintenance​

· One min read
Jean-Philippe Raynaud

High level overview​

The Mithril team created an explorer page with the registered signers to give recognition to the SPOs already running on mainnet. They also worked on refactoring and standardizing errors in the Mithril nodes, and they enhanced the clarity and reduced verbosity of the logs for the signer and aggregator nodes. Additionally, they started working on the implementation of zstandard compression of the snapshot archive, and streaming its download and decompression in the client.

Finally, they implemented machine readable progress information in the client, and streamlined the process of verifying the registration status of a signer.

Low level overview​

  • Completed the issue Create explorer page with registered signers #1097
  • Completed the issue Simplify logs on the Signer/Aggregator #1106
  • Completed the issue Refactor (de)serialization of crypto entities #668
  • Completed the issue Mithril client last snapshot download #1078
  • Completed the issue Use the '/signers/registered' route to verify if a signer is registered #1152
  • Worked on the issue Provide progress information with '--json' option in Client #1095
  • Worked on the issue Use 'zstandard' compression for snapshot archives #876
  • Worked on the issue Make client download and extract the archive simultaneously #1115

· One min read
Franco Testagrossa

High-level summary​

This week, the Hydra team conducted the monthly review meeting for Agust,which continously ensures transparent communication and project evaluation.

In terms of community engagement, the Hydra team gave a Hydra master-class at RareEvo workshop as an educational initiative for the community. Additionally, the team updated the Hydra demo and tutorial to include Mithril and latest release changes.

They also focused on exploring the Hydra network resiliance in order to improve hydra-node availability, as it becomes stuck very easily and that leads to a bad user experience.

What did the team achieve this week​

  • Hydra master-class at RareEvo workshop!
  • Monthly report & review meeting
  • Updated Hydra tutorial including Mithril and re-record demo/tutorial stuff
  • Groomed network resilience and explorer features #188
  • New ADR for “Resourced based API” #1028

What are the goals of next week​

  • Have a poc on network resilience working #188
  • Fix issue #931: “Chain state in head state not updated on replayed observation”
  • Fix issue #1039: "Close transaction dropped from cardano-node"

· 2 min read
Michael Karg

High level summary​

  • Benchmarking: Benchmarking node version 8.2.1 has concluded. Additionally we're developing benchmarking setups for GHC9.6 and UTxO-HD.
  • Infrastructure: Our analysis pipeline has received improvements reducing memory footprint.
  • Tracing: Another batch of optimizations for the new tracing system has been merged; work on namespace consistency guarantees is ongoing.
  • Nomad backend: We're performing and analysing various runs for validation purposes on the new hardware cluster.

Low level overview​

Benchmarking​

We've performed and analysed the benchmarks for the 8.2.1 version of cardano-node as part of our release benchmarking cycle.

Setting up cluster benchmarks requires completing full system integration. This applies to both supporting a new build platform, as is GHC9.6, as well as targeting a specific feature, like a UTxO-HD enabled node. Currently, we're working on respective integrations on both those paths.

Infrastructure​

As cluster runs increase in duration, more and more data is accumulated for analysis. Batch analysis mode needs all data to be held in memory, which wouldn't fit anymore even on a 64GB RAM machine. Changes to the in-memory data representation improving on compactness were able to reduce the RAM requirements of our analysis pipeline.

Tracing​

The next portion of optimizations has been completed and merged to master, getting rid of Haskell's native String representation on critical code paths. This concludes the optimization phase of the new tracing system for all its components used by cardano-node.

The implementation for validation of consistent naming and configuration is ongoing. We're splitting out everything that's verifiable at compile time into a seperate test case which we hope to integrate into CI - leaving only configuration constraints to be verified at or before node startup.

Nomad backend​

The verification phase of the nomad cloud backend is ongoing. We're able to perform full runs on the new hardware cluster and porting profiles and configurations from the legacy one. The goal is to reproduce with confidence known regressions, or improvements, between runs performed on the legacy cluster and runs performed with the new backend.