Prepare IRI release v.1.9.0

[GalRogozinski]

Description

The following features will be included:

• Fix: improve API error response for storeTransactions (improve API error response for storeTransactions #1686)
• Add spent addresses test (Add spent addresses test #1489)
• Fix: Correct walker behaviour for alpha == 0 (Fix: Correct walker behaviour for alpha == 0 #1645)
• Allow to configure benchmark from CLI (Allow to configure benchmark from CLI #1663)
• Bump Jackson Databind to 2.9.10.1 (Bump Jackson Databind to 2.9.10.1 #1677)
• Fix: Don't resave the transaction itself (Fix: Don't resave the transaction itself #1671)
• Fix: Zmq documentation (Fix: Zmq documentation #1665)
• Documentation: Update CONTRIBUTING.MD (Documentation: Update CONTRIBUTING.MD #1631)
• Fix: file descriptors leak in IXI module (Fix file descriptors leak: ScriptException is not the only propagated… #1669)
• Fix: Updated db config loading to account for sections (Fix: Updated db config loading to account for sections #1667)
• Memory optimization for serializing db objects (Memory optimization of StateDiff.bytes() to only use 1 new byte array instead of stateMap.size * 2 arrays #1656)
• Updated RocksDB to 6.3.6 (Updated RocksDB to 6.3.6 #1654)
• Feature: Allow users to tinker with RocksDB configuration via file and upgrade RocksDb (Feature: Allow users to tinker with RocksDB configuration via file; Upgrade RocksDb #1638)
• Feature: Truncate transaction bytes before persisting them to the database (Feature: Truncate transaction bytes before persisting them to the database #1635)
• Update Jdeb plugin
• adds config option to disable sync progress printing (Adds config option to disable sync progress printing #1649)

Also the following PRs should be finalized:
#1694
#1685
#1682 (test failed)
#1660
#1653 (caused memory issues)
#1614
#1584 (need to discuss with devops)
#1561
#1551
#1701 (maybe)

[GalRogozinski]

• fix broken getInclusionStates API call #1685 should be reviewed by @DyrellC
• sort milestone candidates before analyzing #1660 - @acha-bill should fix @galrogo @DyrellC should review
• Fix: Stop submitting zero value bundles as a task into the SpentAddressesService #1614 - @DyrellC should fix
• Feature: Use one RocksDB instance for spent addresses and local snapshots #1584 - See internal checklist
• fix broken getInclusionStates API call #1685 - talk to client guys @galrogo (@jakubcech )
• Feature: Add solidification stage to pipeline and only broadcast solid transactions  #1646 - @DyrellC fix conflict

[GalRogozinski]

Currently waiting for caching and db-size-in-bytes to be merged. Once this is done. The release branch will be created.

The following tests should be made after the release branch was created:

• For caching with high batch rates: pull the plug (stop the node) and see how it recovers.
• For db-size-in-bytes, after another review from @kwek20. Test with different configurations for the DB size, go as small as 1 GB. Test if you can shrink various dbs to the desired size 0 - @DyrellC will do more tests
• For fix broken getInclusionStates API call #1685, manually check the only scenario not covered by the regression test, the status of a non-approved tx that is not a tail. Extra points, if we add it in the regression test. Investigate machine 2 - @DyrellC
• Look at @viossat's issue (Loading DB backup leads to "unexpected error while processing the seen milestones" #1715) from Discord to see if there is a critical bug we can fix on this release.

[GalRogozinski]

• We need to check GTTA performance with Jmeter - @kwek20
• We need to automate the sync test - @karimodm
• Remove configuration recommendation for pruning - @kwek20
• Improve empty LS folder warning If no local snapshot is found create a message that you need to migrate to a new format #1749 - @kwek20
• Change LOCAL_SNAPSHOTS_DB_MAX_SIZE description to explain what -1 means - @kwek20

[GalRogozinski]

In our internal tests:

1. We managed to get to ~35-40 TPS in depth 1
2. We were able to get up to 200 TPS with zmq spamming but Compass didn't approve it all, we need to try it again
3. We saw nodes crash on this high tps but we couldn't crash compass nodes
4. We had syncing problems that should be resolved by Remove ping pong from reply stage #1745

We should run tests again to see where compass breaks but in the meanwhile the results are good enough to create a release branch