Refactoring of Server.h: Isolate server management from other logic

[TTPO100AJIEX]

Logic to manage protocol servers is implemented in the global Server class, which is responsible for the "main" function. I find it reasonable to abstract this in a separate class responsible only for managing the list of servers to improve readability and code encapsulation.

Changelog category (leave one):

• Not for changelog (changelog entry is not required)

[CLAassistant]

All committers have signed the CLA.

[robot-ch-test-poll3]

This is an automated comment for commit 7a313e7 with description of existing statuses. It's updated for the latest CI running

❌ Click here to open a full report in a separate page

Check name	Description	Status
A Sync	There's no description for the check yet, please add it to tests/ci/ci_config.py:CHECK_DESCRIPTIONS	⏳ pending
AST fuzzer	Runs randomly generated queries to catch program errors. The build type is optionally given in parenthesis. If it fails, ask a maintainer for help	❌ failure
CI running	A meta-check that indicates the running CI. Normally, it's in success or pending state. The failed status indicates some problems with the PR	⏳ pending
Integration tests	The integration tests report. In parenthesis the package type is given, and in square brackets are the optional part/total tests	❌ failure
Mergeable Check	Checks if all other necessary checks are successful	⏳ pending
Stateless tests	Runs stateless functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc	❌ failure
Upgrade check	Runs stress tests on server version from last release and then tries to upgrade it to the version from the PR. It checks if the new server can successfully startup without any errors, crashes or sanitizer asserts	❌ failure

Successful checks

Check name	Description	Status
ClickBench	Runs [ClickBench](https://github.com/ClickHouse/ClickBench/) with instant-attach table	✅ success
ClickHouse build check	Builds ClickHouse in various configurations for use in further steps. You have to fix the builds that fail. Build logs often has enough information to fix the error, but you might have to reproduce the failure locally. The cmake options can be found in the build log, grepping for cmake. Use these options and follow the general build process	✅ success
Compatibility check	Checks that clickhouse binary runs on distributions with old libc versions. If it fails, ask a maintainer for help	✅ success
Docker keeper image	The check to build and optionally push the mentioned image to docker hub	✅ success
Docker server image	The check to build and optionally push the mentioned image to docker hub	✅ success
Docs check	Builds and tests the documentation	✅ success
Fast test	Normally this is the first check that is ran for a PR. It builds ClickHouse and runs most of stateless functional tests, omitting some. If it fails, further checks are not started until it is fixed. Look at the report to see which tests fail, then reproduce the failure locally as described here	✅ success
Flaky tests	Checks if new added or modified tests are flaky by running them repeatedly, in parallel, with more randomization. Functional tests are run 100 times with address sanitizer, and additional randomization of thread scheduling. Integration tests are run up to 10 times. If at least once a new test has failed, or was too long, this check will be red. We don't allow flaky tests, read the doc	✅ success
Install packages	Checks that the built packages are installable in a clear environment	✅ success
PR Check	There's no description for the check yet, please add it to tests/ci/ci_config.py:CHECK_DESCRIPTIONS	✅ success
Performance Comparison	Measure changes in query performance. The performance test report is described in detail here. In square brackets are the optional part/total tests	✅ success
Stateful tests	Runs stateful functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc	✅ success
Stress test	Runs stateless functional tests concurrently from several clients to detect concurrency-related errors	✅ success
Style check	Runs a set of checks to keep the code style clean. If some of tests failed, see the related log from the report	✅ success
Unit tests	Runs the unit tests for different release types	✅ success

[rschu1ze]

To be honest, extracting and splitting the code into separate files makes sense if it was used by multiple places but this is not the case here. The only user of "InterServerManager" and "ProtocolServersManager" is Server.cpp. Things are IMHO easier to understand if everything is in one place. I would therefore prefer to not merge this PR (unless there is a good reason which I missed).

[TTPO100AJIEX]

To be honest, extracting and splitting the code into separate files makes sense if it was used by multiple places but this is not the case here. The only user of "InterServerManager" and "ProtocolServersManager" is Server.cpp. Things are IMHO easier to understand if everything is in one place. I would therefore prefer to not merge this PR (unless there is a good reason which I missed).

The idea behind the change was to aid the addition of new protocols - a relatively common improvement that should not require modifications of existing code. Currently, besides implementing the new logic, you would have to change Server.cpp, which is over 2500 lines of code - not the best experience. 
After the change, the code will be encapsulated in ServersManager - a class responsible solely for holding a list of servers, which aids in making the Server responsible for exactly one thing - initialization of all components. With the Manager, everything related to running the servers will remain in one (but different) place explicitly separated from everything else happening in the Server. Thus, to add a new protocol, one would need to modify only a much smaller file - ProtocolServersManager.cpp. 
Furthermore, both InterServersManager and ProtocolServersManager have a hard-coded list of servers to run, which I think is a bad architectural decision. Abstracting this logic in separate classes helps later rework this as well.
Finally, there is similar logic in Keeper.cpp: https://github.com/ClickHouse/ClickHouse/blob/master/programs/keeper/Keeper.cpp#L441 . We might find a way to encapsulate all three places behind a common interface, which improves code readability.

[rschu1ze]

@TTPO100AJIEX Okay, fine with me then, thanks! The only problem I have at this point is that the PR mixes tons of unrelated formatting/cosmetic changes with actual refactorings (moved code). I am really not against formatting changes, but in order to make this reviewable and also revertable (<-- in case something goes wrong), we need to separate the signal from the noise. Therefore, it would be nice if you could rework the commit history (commit messages are at the moment trash anyways - excuse me, just stating the obvious), and make two commits: 1. with all formatting changes 2. the actual refactoring - then force-push and remove "[WIP]" from the PR title.

[TTPO100AJIEX]

@rschu1ze Could you please have a look at the PR? I have removed the styling changes and squashed the commits. There are some failed tests, but I do not understand how the changes could have affected them (especially the upgrade check). Is there a possibility those tests are inherently broken? Could you please clarify, what exactly those tests are checking?

[rschu1ze]

Checking now.

I also restarted the failing tests (they are indeed unrelated).

[rschu1ze]

Thanks! In general, the refactoring looks like it was done right though I did not step through the code or verify it further.

Please see my (mostly minor) comments, after adressing them we can merge.

[TTPO100AJIEX]

@rschu1ze The CI has been completed. Some tests failed, but they seem unrelated to the changes again (and they are different from what had failed before). Could you please merge the PR as it is?

[tavplubix]

@rschu1ze, some CI checks have not finished

[TTPO100AJIEX]

@rschu1ze, some CI checks have not finished

@tavplubix But aren't those checks just broken or flaky? Something fails for all PRs, and the changes do not seem to be related to the failed tests.
Are any more modifications required to merge this change?

[rschu1ze]

@TTPO100AJIEX As you probably know, ClickHouse also maintains an internal repository and all "public" PRs are picked in the background into the non-public repo. In this case, there was a merge conflict in the non-public repository (as can be expected by such large refactorings) but I merged your PR yesterday without noticing. By bad, sorry. I'll do a revert of the revert soon and check/resolve these conflicts.

[javisantana]

I do think you guys should give a heads up about this kind of things to the community, otherwise it'll be impossible to contribute in a few months (when your fork diverges enough)

[rschu1ze]

Sorry for the delay, priorities got a bit disarranged for me this week. I'll definitely check today, promised.