HORNET-0.5.1-rc1

Added

- webapi route to control the spammer plugin
- getSpammerTips webapi call

Changed

- findTransaction now returns an intersection of the search criteria
- show database size in GB if size > 1 GB (dashboard)
- moved dashboard frontend code to another git repo (https://github.com/iotaledger/node-dashboard)
- spammer doesn't start automatically at node startup
- getNodeInfo now returns the connected peers count

Fixed

- OOM bug in the future cone solidifier
- map concurrent write/read panic in websocket
- warpsync milestone requesting
- re-add logger settings to mainnet config file

Config file changes

config.json

  "httpAPI": {
+    "permittedRoutes": [
+      "healthz"
+    ],
  }
+  "logger": {
+    "level": "info",
+    "disableCaller": true,
+    "encoding": "console",
+    "outputPaths": [
+      "stdout"
+    ]
+  },
  "spammer": {
-    "semiLazyTipsLimit": 30
  },

HORNET-0.5.0

⚠️ Breaking change ⚠️

The old HORNET versions won't be functional within the mainnet anymore!

Upgrading to 0.5.0 from a previous version:

1. The new coordinator address is: UDYXTZBE9GZGPM9SSQV9LTZNDLJIZMPUVVXYXFYVBLIEUHLSEWFTKZZLXYRHHWVQV9MNNX9KZC9D9UZWZ. We suggest that you simply take the config.json from the repository and change it according to your needs. The new coordinator address is the default value in the node, so it isn't defined in the config. If you're taking over your current config.json (which we don't recommend since a lot of keys changed), make sure to change the coordinator options to the new values.
2. Start up your v0.5.0 node with the --overwriteCooAddress and let it synchronize.
3. At this point you can let your node run but eventually you should restart it without the --overwriteCooAddress flag.

Added

- White-Flag confirmation
- Weighted uniform random tipselection for nodes
- Adaptive heaviest branch tipselection for coordinator
- Optional powsrv.io PoW support
- LMI and neighbor counts to dashboard
- getTipInfo API call
- "isHealthy" to getNodeInfo
- conf_trytes (confirmed trytes) ZMQ topic
- conf_trytes (confirmed trytes) MQTT topic
- Conflicting badge to the transaction explorer
- Add autopeering rule to drop neighbors with LSMI below our pruning index
- Automatic dashboard websocket reconnect
- Database tainted flag for coordinator

Changed

- Request tx from all neighbors that could have the data
- Bump protocol feature set for whiteflag (breaking protocol change)
- Reduced dashboard traffic by introducing subscriptions to topics
- Store binary trunk, branch, and bundle hashes in metadata to reduce load on IO
- Improve caching strategy in solidification and confirmation
- Improve caching in the dag helpers
- Reduce recursion in the future cone solidifier
- Use stack based DFS in TraverseApprovees
- Use stack based BFS in TraverseApprovers
- Return false for conflicting tx in the getInclusionStates web api call
- Coordinator now waits until the milestone is solid
- Set higher default verticesLimit in visualizer
- Update to Go 1.15

Removed

- Unused RefsInvalidBundles cache

Fixed

- Database revalidation
- Missing byte to trytes conversion in some error messages
- Make code more testable
- Adding autopeered neighbors as static neighbors
- CTPS calculation
- Spikes in conf.rate calculation
- Pruning
- Do not drop autopeering and "acceptAny" peers on peering.json change
- Index out of range in attachToTangle
- Deadlock in snapshotting and pruning
- Error reason for connection abort is now shown
- Peering configs now recognized via CLI
- Coordinator bootstrapping
- Milestone missing / Milestone updated panics
- Hash conflicts in the visualizer
- SupportedFeatureSets logic in handshake
- Websocket/dashboard deadlock
- Dashboard visualizer re-rendered to often
- Do not panic if snapshot creation is aborted
- Fix solid entry point indexes

Removed

- Unused defaults from config.json
- Graph plugin
- Monitor plugin
- Legacy gossip protocol
- Genesis tx special case

Config file changes

Please use the new config.json and transfer values from your current config.json over to the new one, as a lot of keys have changed or got removed (instead of mutating your current one).

HORNET-0.5.0-rc13

Breaking change:
** DO NOT USE IT ON MAINNET, IT WILL CRASH IMMEDIATELY AND IT WILL DESTROY YOUR DATABASE !!! **

Fixed

- Fix TransactionConfirmed event

HORNET-0.5.0-rc12

Breaking change:
** DO NOT USE IT ON MAINNET, IT WILL CRASH IMMEDIATELY AND IT WILL DESTROY YOUR DATABASE !!! **

Added

- Add isHealthy to getNodeInfo
- Database tainted flag for coordinator
- Add additional security check in whiteflag

Changed

- Store binary trunk, branch, and bundle hashes in metadata to reduce load on IO
- Improve caching strategy in solidification and confirmation
- Improve caching in the dag helpers
- Reduce recursion in the future cone solidifier
- Disable autopeer stale check by default
- Update test pipeline to Go 1.15

Fixed

- Fix deadlock at object storage flush (revalidation and shutdown)
- Do not panic if snapshot creation is aborted
- Fix solid entry point indexes
- Fix spikes in conf.rate calculation

HORNET-0.5.0-rc11

Breaking change:
** DO NOT USE IT ON MAINNET, IT WILL CRASH IMMEDIATELY AND IT WILL DESTROY YOUR DATABASE !!! **

Changed

- Remove tanglebay snapshot source for comnet

Fixed

- Fix deadlock in revalidation
- Fix possible deadlock at coordinator shutdown
- Fix panic in dashboard websocket handling

HORNET-0.5.0-rc10

Breaking change:
** DO NOT USE IT ON MAINNET, IT WILL CRASH IMMEDIATELY AND IT WILL DESTROY YOUR DATABASE !!! **

Added

- checkConsistency webapi call for compatibility reasons
- getTipInfo webapi call to signal if transactions need to be promoted or reattached
- Remove stale autopeers if dropped packages exceed a defined percentage of total packets

Changed

- Remove unused defaults from config.json

Fixed

- Dashboard misc page not shown if spammer activated
- Dashboard websocket reconnected too often
- Panic in URTS UpdateScores if the node was unsync for a longer time

HORNET-0.5.0-rc9

Breaking change:
** DO NOT USE IT ON MAINNET, IT WILL CRASH IMMEDIATELY AND IT WILL DESTROY YOUR DATABASE !!! **

Added

- Automatic dashboard websocket reconnect

Changed

- Reduced dashboard traffic by introducing subscriptions to topics

Fixed

- Websocket/dashboard deadlock
- Dashboard visualizer re-rendered to often
- Database revalidation

HORNET-0.5.0-rc8

Breaking change:
** DO NOT USE IT ON MAINNET, IT WILL CRASH IMMEDIATELY AND IT WILL DESTROY YOUR DATABASE !!! **

Added

- Latest solid milestone and neighbor counts of every neighbor to dashboard
- powsrv.io support

Changed

- Breaking protocol version change due to whiteflag
- Heartbeat message was extended by neighbors count and latest solid milestone
- Requests are now sent if an unsynced neighbor may also have the data based on latest solid milestone

Fixed

- SupportedFeatureSets logic in handshake
- Milestone missing / Milestone updated panics
- Database revalidation
- Autopeered neighbors can now be added as static neighbors
- Hash conflicts in the visualizer
- Wrong CTPS calculation on bundle tail reattachments

HORNET-0.5.0-rc7

Breaking change:
** DO NOT USE IT ON MAINNET, IT WILL CRASH IMMEDIATELY AND IT WILL DESTROY YOUR DATABASE !!! **

Changed

- URTS now uses separate tip-pools for semi- and non-lazy to prevent below-max-depth attacks
- Coordinator now waits until the milestone is solid
- Definition of semi-lazy tips in URTS was changed
- Spammer now tries to reduce semi-lazy tips if certain threshold reached
- Use dag helper in whiteflag (#583)

Fixed

- Coordinator bootstrapping
- Peering configs now recognized via CLI
- Error reason for connection abort is now shown
- Deadlock in snapshotting and pruning
- Fix wrong approveeHashes in URTS calculateScore

HORNET-0.5.0-rc6

Breaking change:
** DO NOT USE IT ON MAINNET, IT WILL CRASH IMMEDIATELY AND IT WILL DESTROY YOUR DATABASE !!! **

Added

- Added whiteflag test setup (#572)
- Add autopeering rule to drop neighbors with LSMI below our pruning index

Removed

- Unused RefsInvalidBundles cache

Fixed

- Visualizer: wrong color of some connections between transactions
- Index out of range in attachToTangle
- Do not drop autopeering and "acceptAny" peers on peering.json change