Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[v24.1.x] Fix ghost node handling in raft heartbeats #23934

Merged
merged 3 commits into from
Oct 29, 2024
Merged

[v24.1.x] Fix ghost node handling in raft heartbeats #23934

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
be49257
tests: decom before waiting for cluster health in admin_uuid_operatio…
ztlpn Oct 28, 2024
44ee1d7
raft: reply to heartbeats with group_unavailable if addressed to wron…
ztlpn Oct 25, 2024
cb8a4c0
raft: ignore heartbeat replies from unexpected node ids
ztlpn Oct 25, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions src/v/raft/consensus.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3930,7 +3930,7 @@ reply_result consensus::lightweight_heartbeat(
target_node,
_self,
source_node);
return reply_result::failure;
return reply_result::group_unavailable;
}

/**
Expand Down Expand Up @@ -3985,7 +3985,7 @@ ss::future<full_heartbeat_reply> consensus::full_heartbeat(
target_vnode,
_self,
source_vnode);
reply.result = reply_result::failure;
reply.result = reply_result::group_unavailable;
co_return reply;
}
/**
Expand Down
11 changes: 11 additions & 0 deletions src/v/raft/heartbeat_manager.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -455,6 +455,17 @@ void heartbeat_manager::process_reply(
return;
}
auto& reply = r.value();

if (reply.source() != n) {
vlog(
raftlog.warn,
"got heartbeat reply from a different node id {} (expected {}), "
"ignoring",
reply.source(),
n);
return;
}

reply.for_each_lw_reply([this, n, target = reply.target(), &groups](
group_id group, reply_result result) {
auto it = _consensus_groups.find(group);
Expand Down
17 changes: 6 additions & 11 deletions tests/rptest/tests/admin_uuid_operations_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -276,14 +276,12 @@ def test_force_uuid_override(self, mode):
backoff_sec=2,
err_msg=f"{to_stop.name} did not take the UUID override")

self.logger.debug(f"Wait for the cluster to become healthy...")
self.logger.debug(f"Decommission ghost node [{ghost_node_id}]...")
self._decommission(ghost_node_id)

self.logger.debug(f"...and wait for the cluster to become healthy.")
self.wait_until_cluster_healthy(timeout_sec=30)

self.logger.debug(
f".. and decommission ghost node [{ghost_node_id}]...")
self._decommission(ghost_node_id)

self.logger.debug(
"Check that all this state sticks across a rolling restart")

Expand Down Expand Up @@ -373,14 +371,11 @@ def test_force_uuid_override_multinode(self, mode):
auto_assign_node_id=True,
)

self.logger.debug("Wait for the cluster to become healthy...")
self.logger.debug(f"Decommission ghost node [{ghost_node_id}]...")
self._decommission(ghost_node_id)

self.logger.debug("...and wait for the cluster to become healthy.")
controller_leader = self.wait_until_cluster_healthy(timeout_sec=30)

assert controller_leader is not None, "Didn't elect a controller leader"
assert controller_leader not in to_stop, f"Unexpected controller leader {controller_leader.account.hostname}"

self.logger.debug(
f"...and decommission ghost node [{ghost_node_id}]...")

self._decommission(ghost_node_id)
Loading