Rollback version patching mechanism - Closes #2466

[jondubois]

What was the problem?

Our approach to versioning should be compatible with older nodes.
A previous incomplete patch needed to be rolled back. We need to change our approach to versioning to be compatible with testnet nodes.

How did I fix it?

• Reverted commits in Update P2P layer connection logic so it can work with 1.0.0-rc.* releases #2389 and https://github.com/LiskHQ/lisk/pull/2444/commits
• Changed the version number in package.json to 1.1.0
• Fixed an issue with integration tests which was causing it to count the same node monitoring sockets multiple times after a node restarted.

How to test it?

Run the node and connect to testnet, peer disconnects with code 1000 should not (or rarely) happen during the sync phase.

Review checklist

• The PR resolves Rollback version patching mechanism #2466
• All new code is covered with unit tests
• All new code was formatted with Prettier
• Linting passes
• Tests pass
• Commit messages follow the commit guidelines
• Documentation has been added/updated

[webmaster128]

After this change, 1.1.0-rc.0 can still not connect to 1.0.0-rc.4 and 1.0.0-rc.5 nodes. I needed to revert be4d8fd in order to get current 1.1.0 branch synchonizing.

By the way: why minVersion bump again? This is the next hard fork

[jondubois]

@webmaster128 That's correct. This PR reverts a patch which did not work as intended. There were two options; push forward with the hack (making more changes to the code) until everything appears to work, or roll back what we had and find an alternative solution.

Since this issue only affects testnet, it was decided that the best solution for it did not require a code change. There should be an announcement very soon about next steps for testnet 1.1.

[webmaster128]

@jondubois thanks for the info. I am happy this is expeted behaviour and well known. Looking forward to see 1.1 being deployed.

[MaciejBaj]

@webmaster128 I summarized the steps here #2389 (comment).