Release v0.8.18

This release fixes an issue where the node may consider a block sent in the future as a valid block (see #1942). Stake pool operators are recommended to upgrade to this version.

Release v0.8.17

This release contains mostly new features and enhancements. On the REST APIs: it is now possible to access more data from the startup point, allowing the users to follow progressions of the node updating with the network. On the p2p:

• the first one is a whitelist has been implemented to prevent the policy to quarantine specifics IPs #1982
• The second one is a new layer: the preferred list. #1990

This preferred list is interesting for nodes to team up in the network in order to establish inter-connections faster. For example this is useful for operators maintaining multiple nodes (like IOHK's). see documentation for configuration options.

Detailed Changelogs

Release v0.8.16

Some minor fixes in the latest added endpoints regarding stake distribution and reward history. The node was throwing an exception when the epoch didn't exist. Now it returns a proper 404 HTTP code, this will make client apps easier to integrate.

Also we have added CPU optimised assets to build on future releases.

Detailed Changelogs

Release v0.8.15

Many improvements on the network:

• updates on the P2P policy, increasing the time a node is quarantined if recidivists;
• new field in the stats REST APIs: peerConnectedCnt;
• refine the default values of the max_client_connections;
• core improvement in the blockchain state engine;

This release has received an ever growing number of contributions from the community, thanks <3

Detailed Changelogs

Release v0.8.14

Many core changes and new features were added. One of the most wanted feature is to be able to retire a stake pool. The code to do so was always there but it was not easily possible to retire a stake pool. Now jcli exposes a command line option to create a stake pool retirement certificate.

Another important feature is to be able to access via REST APIs the Stake Distribution of an epoch and the reward distribution history.

Notable changes:

• add stake pool retirement to jcli #1809
• access the stake distribution of any epoch #1901
• access the reward history of any epoch #1886
• add the parent block hash of the minted block in the leadership log #1883

Also note that you need --rewards-report-all in order to see all the reward distribution for all accounts in #1886 or when using JORMUNGANDR_REWARD_DUMP_DIRECTORY.

Detailed Changelogs

Release v0.8.13

Fix a regression that came with 0.8.12 in the REST APIs for the leader handling (#1857)

Detailed Changelogs

v0.8.12

Improve bootstrap reliability, speed and verbosity, and some other quality of life improvements

• Improve the speed of loading from storage and the general streaming from storage
• Netboot is now a bit more comfortly verbose, adding some location and speed
• Fix peer bootstrap to timeout properly when the peer doesn't play ball
• Improve memory bound of the pending/approved fragments
• Cleaning more code by using latest dependencies, std futures, and newer async dependencies
• Enclave now detect having duplicated leaders loaded and creating blocks competitively

v0.8.11

Some improvements in the bootstrap, and network stability

The node has significant more flexibility on the bootstrapping process; It can now bootstrap the larger than normal block0 from a HTTP url, do trusted local startup by using --genesis-block0-hash together with --genesis-block0, and now default to bootstrap from a more diverse set of peers which is received from the trusted peers, although we kept the old behaviour with the config stanza bootstrap_from_trusted_peers.

The node now requires the trusted peers to not be empty (it will be a hard error soon), and there is a skip_bootstrap configuration field that allow to run as a self-node or allow what people have been doing by removing the trusted peers, although we hope it's not necessary anymore.

Poldercast is now updated 0.11.3 which set some limits to the general size of what we know about the network to prevent wheel spinning on a unbounded size of gossip, and the code has made progress toward integrating 0.12.

chain-libs has been updated, and now pool parameters update will block fees changes, although we hope to restore fee reductions as soon as possible, and maybe start a dialogue on how to properly do fee increasing also.

Finally the fragment subsystem is now lockless and more efficient, one step further toward making sure we don't lock other part of the system, or introduce long delays around block creation.

All the new configuration options, including the a working block0 fetcher using github:

http_fetch_block0_service:
- "https://github.com/input-output-hk/jormungandr-block0/raw/master/data/"
skip_bootstrap: false
bootstrap_from_trusted_peers: false

Full Changelog

v0.8.10-2

build fixes for cross compilation targets (arm, aarch64, musl, ..) and other CI release compilation fixes due to the change from travis/appveyor to github

v0.8.9

Fix v0.8.8 issues with bootstrapping

• fix panic in network bootstrap where block where ahead of the validated ones
• use checkpoints when bootstrapping

Full Changelog