v0.8.8

Some minor improvements to logging, bootstrapping and some temporary relief fixes for cpu/locking pains

• Improvements to logging in storage bootstrap, network tasks and chain-selection
• Improve bootstrapping reliability by retrying
• Reduce the cpu impact of poldercast algorithm with a temporary fix
• Reduce the issues of synchronous locking on the asynchronous stack while we carry on with
  the migration to asynchronous

Full Changelog

v0.8.7

Another release with minor but steady network improvements which should lead
to a more stable blockchain network, and some technical debt reductions.

This should lead to improvements on the overall connections management, so that
an oversubscribed node has always room to open a new connection for its own
benefits and that idle connections are recycled more aggresively. Some
improvements on the locking will also reduce performance problems.

We also have some dependencies updates, and some parts has been rewritten to
the now rust standard async and futures, leading to sustantially simpler
and cleaner code in many places. This is just the start of this work.

Full changelog

v0.8.6

Summary

New 0.8.6 release aims to improve greatly the network stability of the node. But not only. There have been a lot of code quality improvement in the network code, the storage and the service manager (the module that monitor all the tasks running in the node).

This release does not provide a quick solution to some of the issues reported by the community such as the clone stake pools problem. This is an issue linked to the incentive model and the issues of the network. Improving the network has a higher priority right now.

Detailed Changelog

Full Changelog

Implemented enhancements

• Get rid of CandidateForest #1569
• Update chain-deps for rand and network API changes #1507
• Implement a TCP server connection manager that rejects incoming connections when full #1497

Fixed bugs

• android compilation fails with rlimit #1553
• SqliteFailure: "no such table: Blocks" #1485
• Tests are sporadically failing on Circle CI #1463
• 0.8.5-alpha1 panics observed #1422
• no data at API endpoint /api/v0/stake_pool #1421
• Error in documentation of public_id #1420
• disable diagnostic on Android #1557
• replace task key with subtask for topology policy #1533
• enable shared cache for in-memory databases #1508
• enforce the correct order of writes to storage #1441

Closed issues

• thread 'tokio-runtime-worker-1' panicked at 'called Option::unwrap\(\) #1564
• Upgrade to Async/Await: use tokio compat Runtime to allow new async/await use #1548
• Upgrade to Async/Await: use default runtime #1547
• CRIT task panicked #1542
• Cannot sign the block: This leader 1 is not in the enclave #1540
• Error processing incoming header stream -Mac OS Mojave #1535
• "Block theft" seems to be happening #1532
• Jormungandr - logs - poldercast / policy log task / sub_task #1499
• restart of SSH terminal causes node to stop syncing #1480
• Blockchain is not moving up - jormungandr v0.8.5 #1479
• Run jormungandr with version testnet Byron #1477
• Ghost block #1464
• CRIT Task panicked, task: leadership #1451
• v0.8.5 labeled as a Pre-release #1450
• stats: Server Error: 500 Internal Server Error #1449
• Leader log should have a flag if block was invalidated #1446
• Dec 25 00:51:54.086 WARN blockchain is not moving up, the last block was 4963 seconds ago, task: stuck_notifier #1443
• The Node is Not in Sync v0.8.5 #1440
• Node crash #1434
• itn_rewards_v1 - jcli generates a binary genesis which hash is different from the official one #1430
• Block produced but is not visible in Cardano Explorer #1427
• the node is not synced v0.8.3 #1410
• CRIT Task panicked, task: block #1408
• ERRO cannot compute the time passed #1404
• Wrong parent block - j. 0.8.3 #1403
• The lastBlockTime increases but lastBlockHeight freezes #1402
• Error: deadline has elapsed #1392

Merged pull requests

• Revert "use a global runtime instead of a local one and new future" #1570
• remove unused shell.nix file #1559
• [Tests] Disruption private network test #1558
• use a global runtime instead of a local one and new future #1556
• Dockerfile uses VER as an ARG but not VERSION #1555
• miss to cp a shell script #1554
• integrate new SQLiteBlockStore constructors #1550
• use with\_executor as per breaking changes from chain-libs/network-grpc #1549
• use one Runtime for all the services. #1546
• Remove some dead code #1545
• Make server connection failures non-fatal #1531
• rearrange sections, simplify text #1528
• [Tests] Jts soak test #1525
• Thread pool for server tasks #1523
• add RELEASE file #1520
• [Tests] Reward test fix #1516
• [Test] Implement long soak test (Selfnode) #1515
• Docs: configuration/network - gossip_interval - fix default value #1506
• [Tests] NodeStats struct update #1487
• Command for generating the public id #1486
• Clean up docs for network public_id #1484
• Fix stake pool OpenAPI lacking pool_id parameter #1483
• [Test] Testnet: added "Port already in use" new error code #1482
• [Tests] build fix for explorer pr #1481
• [Test] Collect reward test fix #1478
• [Tests] Explorer test #1414

v0.8.5

New stable version with different bug fixes and enhancements (see alpha1, 2 and 3).

Implemented enhancements:

• leverage SQLite multi-threading power by removing the bottleneck lock #1412

Fixed bugs:

• Fix block time issue in the REST API and documentation #1426
• Handle concurrency in CandidateForest::apply_block #1425
• Thread 'blockX' panicked at missed chain pull root candidate #1388
• blockchain: More robust CandidateForest #1405

Closed issues:

• CRIT Service has terminated with an error #1418

v0.8.5-alpha3

Fixed bugs

• Fix block time issue in the REST API and documentation #1426

v0.8.5-alpha2

Fixed bugs:

• Handle concurrency in CandidateForest::apply_block #1425

Closed issues:

• CRIT Service has terminated with an error #1418

v0.8.5-alpha1

early version for testing only

Implemented enhancements

• leverage SQLite multi-threading power by removing the bottleneck lock #1412

Fixed bugs

• Thread 'blockX' panicked at missed chain pull root candidate #1388
• blockchain: More robust CandidateForest #1405

Merged pull requests

• [Tests] Stability fix - Change reward test parameters #1393

v0.8.4

This new release fixes 2 bugs in the explorer GraphQL backend and is a step toward slightly improved network stability via better handling of concurrent pull requests from the network.

Fixed bugs

• Panic: transaction not found for utxo input #1382
• Panic on 'invalid key' #1381
• Question: initial bootstrap failed, error: BlockMissingParent... referring to block0, but wrong parent. #1326
• Eliminate duplicates in transactions by address #1397
• handle utxo pointers in the same block #1394
• Fix a race condition of header pull with block pull #1389
• Exclude block 0 from sent branches #1385

Closed issues

• peer node ID differs from the expected X hash, node id: Y hash #1390
• Compilation error building latest master on MacOS #1376
• Limit concurrent pull requests #1349

Merged pull requests

• Limit concurrent pull requests #1365
• remove custom_error #1345

v0.8.3

This release contains mainly a bug fix on the network. Improving how the network react with concurrent chain pull and preventing some cases of invalid concurrency.

Fixed bugs:

• Node stuck - block is already cached as a candidate - panic 'assertion failed: _old.is_none()' - immediate memory increase #1327
• jcli transaction make-witness fails with error #1323

Closed issues:

• jcli rest v0 stake get no longer lists all pools with delegated stake in 0.8.2 #1371
• where is jcli? #1369
• Error in the overall configuration of the node/ bootstrap file. v0.8.2 #1363
• 404 Not Found status code is too generic for unused addresses which confuses jcli users #1361

Merged pull requests:

• [Tests] Another Fix for non functional tests #1378
• [Tests] fixed non_functional tests compilation issues #1375
• Reverts change in 7e79334 which remo… #1362
• Docs: jcli/transaction + stake_pool/delegating_stake - fix #1347
• Rework chain pull to be concurrency friendly #1346

v0.8.2

Bugfix:

• accounting issue: fix a fee accounting issue when using owner stake delegation certificate, where fees were lost from the ledger.