Fix multi DC token calculation

[dkropachev]

Provision multi DC cluster and list initial_token:

rm -rf ~/.ccm/test
ccm create test -i 127.0.0. -n 4:0 -v release:6.2 --scylla
cat ~/.ccm/test/*/node.conf | grep init
initial_token: -9223372036854775808
initial_token: -9223372036854775708

As you can see only two nodes got initial_token
To fix that we need to pass node_locations to balanced_tokens_across_dcs And have ti fixed a bit.

After the fix:

rm -rf ~/.ccm/test
ccm create test -i 127.0.0. -n 4:0 -v release:6.2 --scylla
cat ~/.ccm/test/*/node.conf | grep init
Current cluster is now: test
initial_token: -9223372036854775808
initial_token: -4611686018427387904
initial_token: 0
initial_token: 4611686018427387904

Fixes: #642

[dkropachev]

@fruch , @Bouncheck

[fruch]

1. not sure how one of the nodes has 0, I don't understand the logic here
2. we should run dtest gating to validate it's not breaking (not that I think any dtest is using this option)

EDIT: there a few test that do use it: https://github.com/search?q=repo%3Ascylladb%2Fscylla-dtest%20initial_token&type=code

[dkropachev]

@fruch , isreal, can you please approve the pipeline

[dkropachev]

1. not sure how one of the nodes has 0, I don't understand the logic here
2. we should run dtest gating to validate it's not breaking (not that I think any dtest is using this option)

EDIT: there a few test that do use it: https://github.com/search?q=repo%3Ascylladb%2Fscylla-dtest%20initial_token&type=code

Running: https://jenkins.scylladb.com/view/staging/job/scylla-staging/job/dkropachev/job/dtest-gating/2

[dkropachev]

@fruch , Two runs failed at the same test:

        if found_errors:
            logger.error(f"Unexpected errors found: {found_errors}")
            errors_summary = "\n".join([f"{node}: {len(errors)} errors\n" + "\n".join(errors[:5]) for node, errors in found_errors])
>           raise AssertionError(f"Unexpected errors found:\n{errors_summary}")
E           AssertionError: Unexpected errors found:
E           node1: 1 errors
E           ERROR 2025-03-18 16:01:01,423 [shard 0:main] seastar_memory - seastar/src/core/memory.cc:2126 @void seastar::memory::maybe_dump_memory_diagnostics(size_t, bool): failed to log message: fmt='Failed to allocate {} bytes at {}': std::bad_alloc (std::bad_alloc)

It is known problem, we can consider it passed.

[fruch]

LGTM