Georgy Moshkin authored 2 years ago and Georgy Moshkin committed 2 years ago

deactivation means that instance
- is demoted to learner (if it wasn't one already)
- is marked "inactive" so that it's ignored when determining
    the number of voters required for the cluster

Georgy Moshkin authored 2 years ago and Georgy Moshkin committed 2 years ago

Topology::diff, Topology::to_replace, etc. only store ids, so that
peer info is consistent

Yaroslav Dynnikov authored 2 years ago and Yaroslav Dynnikov committed 2 years ago

Bootstrapping the replication was implemented in 2dac77c5. But it was
configured on the new instance only. The old instance (that joined
earlier) couldn't update `box.cfg({replication})` until now.

Close https://git.picodata.io/picodata/picodata/picodata/-/issues/52

Just supply it with a default value "demo". No new tests are necessary,
we already have `test/int/test_joining.py::test_cluster_id_mismatch`.

Close https://git.picodata.io/picodata/picodata/picodata/-/issues/96

If proc_discover is invoked after raft node was initialized but before
raft leader was elected, it would return an error before this commit.
Because of that it was impossible to restart the whole cluster at once.

This commit change proc_discover such that in case leader_id is not
ready, the normal discovery algorithm takes place.

Closes #93

Georgy Moshkin authored 2 years ago and Georgy Moshkin committed 2 years ago

Was broken because `tarantool_free` checks if the current process is the
main one and not the child, which was forked at some point (at what
point?). This check was implemented by saving the original process's id
in the static variable master_pid, which got set when the code got
loaded the first time into memory.

So we broke this when we started forking the process in picodata, which
resulted in `master_pid` being set to the pid of the picodata's
"supervisor" process, which doesn't even enter tarantool runtime.

Closes #37

Restarting both instances doesn't work yet, to be fixed later.

Close https://git.picodata.io/picodata/picodata/picodata/-/issues/90

Since commit d87dd4ca `leader_id` became an `Option`, so the `None`
value isn't rendered in the `picolib.raft_status` response:

```python
status={'is_ready': False, 'raft_state': 'Follower', 'id': 1}
```

It makes pytest complain about missing argument:

```
cluster2 = Cluster("127.0.0.1:3300", n=2)

    def test_restart_leader(cluster2: Cluster):
        i1, i2 = cluster2.instances
        i1.assert_raft_status('Leader')
        i2.assert_raft_status('Follower')

        i1.restart()
>       i1.wait_ready()

test/int/test_joining.py:209:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
../../.local/share/virtualenvs/picodata-6sv6l6y-/lib/python3.10/site-packages/funcy/decorators.py:45:
in wrapper
    return deco(call, *dargs, **dkwargs)
../../.local/share/virtualenvs/picodata-6sv6l6y-/lib/python3.10/site-packages/funcy/flow.py:127:
in retry
    return call()
../../.local/share/virtualenvs/picodata-6sv6l6y-/lib/python3.10/site-packages/funcy/decorators.py:66:
in __call__
    return self._func(*self._args, **self._kwargs)
test/int/conftest.py:305: in wait_ready
    status = self._raft_status()
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

self = Instance(i1, listen=127.0.0.1:3301)

    def _raft_status(self) -> RaftStatus:
        status = self.call("picolib.raft_status")
        assert isinstance(status, dict)
        eprint(f"{status=}")
>       return RaftStatus(**status)
E       TypeError: RaftStatus.__init__() missing 1 required positional
argument: 'leader_id'

test/int/conftest.py:280: TypeError
```

This patch fixes the failure message:

```
self = Instance(i1, listen=127.0.0.1:3301)

    @funcy.retry(tries=20, timeout=0.1)
    def wait_ready(self):
        status = self._raft_status()
>       assert status.is_ready
E       AssertionError: assert False
E        +  where False = RaftStatus(id=1, raft_state='Follower',
is_ready=False, leader_id=None).is_ready

test/int/conftest.py:306: AssertionError
```

* Make `--cluster-id` CLI mandatory.
* Handle cluster_id mismatch in raft_join.
  When an instance attempts to join the cluster and the instances's
  `--instance-id` parameter mismatches the cluster_id of the cluster
  an error is raised inside the raft_join handler.

Previously the discovery algorithm would try to reach each known peer
sequentially requiring each consequent request to succeed until the next
one can be attempted. This would not work in some cases (see test in
previous commit).

So the new algorithm instead makes a single attempt to reach each peer
within a round, and if some failed they're retried in the next round of
requests. This allows overall discovery to succeed in cases when some
of the initial peers never respond.

Closes #54