github ava-labs/avalanchego v1.2.1
Apricot Phase 0 - Upgrade 1 - Patch 1

latest releases: v1.12.0-initial-poc.9, v1.12.0-initial-poc.8, v1.12.0-initial-poc.7...
3 years ago

This update is backwards compatible. It is optional, but encouraged. The patch includes stability, performance, and monitoring improvements. Note that this update removes network-timeout-increase and network-timeout-reduction as command line arguments.

Change summary:

  • Added UTXOs to the platformvm.getStake response.
  • Added benchlist reporting to the info.peers response.
  • Added additional health checks into the networking layer.
  • Added percent of stake connected as a reported metric.
  • Added bootstrapping restart logic to ensure the node has caught up to the current tip, even during times of high throughput.
  • Added subnet-wide bootstrapping to ensure that a chain won't fall behind due to another chain bootstrapping.
  • Prevented verification of rejected blocks to avoid unnecessary computation.
  • Removed gossiping of non-preferred blocks to the network.
  • Switched the network timeout calculator to use an EWMA of the observed network latency.
  • Removed Get requests from the network latency calculations.
  • Cleaned up the benchlisting algorithm.
  • Cleaned up handling of dropped messages on send.
  • Cleaned up outstanding request and timeout logic.
  • Generalized performance tracking to allow for prefixing of profile names.
  • Added additional caching to the Avalanche bootstrapping traversal.
  • Fixed ansible linting.

The added command line arguments mainly consist of configurations of health checks. Additionally, the modified network latency calculations changed the name of some command line args.

Added command line arguments:

  • network-timeout-halflife
  • network-timeout-coefficient
  • network-health-min-conn-peers
  • network-health-max-time-since-msg-received
  • network-health-max-time-since-msg-sent
  • network-health-max-portion-send-queue-full
  • network-health-max-send-fail-rate
  • network-health-max-time-since-no-requests
  • router-health-max-drop-rate
  • router-health-max-outstanding-requests
  • health-check-frequency
  • health-check-averager-halflife
  • bootstrap-retry-enabled
  • bootstrap-retry-max-attempts

Removed command line arguments:

  • network-timeout-increase
  • network-timeout-reduction

Don't miss a new avalanchego release

NewReleases is sending notifications on new releases.