mirrors/scylladb
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

tablets, topology_coordinator: Skip tablet draining on replace

Browse Source 
Replace doesn't drain (rebuild) tablets during topology change. They
are rebuilt afterwards when the replaced node is in "left" state and
replacing node is in normal state. So there is no point in attempting
to drain, as nothing will be drained.

Not only that, doing so has a risk, because the load balancer is
invoked on a transitional topology state in which we can end up with
no normal nodes in a rack. That's the case if the replaced node was
the last one in the rack. This tripped one of the algorithms which
computes rack's shard count for the purpose of determining ideal
tablet count, it was not prepared to find an empty rack to which a
table is still repliacated. That was fixed separately, but to avoid
this, we better skip tablet draining here.
This commit is contained in:
Tomasz Grabiec
2025-10-28 18:17:48 +01:00
parent 5e03b01107
commit 288e75fe22
1 changed files with 5 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
service/topology_coordinator.cc
View File

@@ -2325,12 +2325,14 @@ class topology_coordinator : public endpoint_lifecycle_subscriber {
break;
}
builder.set_transition_state(topology::transition_state::tablet_draining)
guard = take_guard(std::move(node));
builder.set_transition_state(topology::transition_state::write_both_read_old)
.set_version(_topo_sm._topology.version + 1)
.set_session(session_id(guard.new_group0_state_id()))
.with_node(node.id)
.set("tokens", it->second.ring->tokens);
co_await update_topology_state(take_guard(std::move(node)), {builder.build()},
"replace: transition to tablet_draining and take ownership of the replaced node's tokens");
co_await update_topology_state(std::move(guard), {builder.build()},
"replace: transition to write_both_read_old and take ownership of the replaced node's tokens");
}
break;
default: