Files
scylladb/test/boost/storage_proxy_test.cc
Calle Wilund f18e967939 storage_proxy: Make split_stats resilient to being called from different scheduling group
Fixes #11017

When doing writes, storage proxy creates types deriving from abstract_write_response_handler.
These are created in the various scheduling groups executing the write inducing code. They
pick up a group-local reference to the various metrics used by SP. Normally all code
using (and esp. modifying) these metrics are executed in the same scheduling group.
However, if gossip sees a node go down, it will notify listeners, which eventually
calls get_ep_stat and register_metrics.
This code (before this patch) uses _active_ scheduling group to eventually add
metrics, using a local dict as guard against double regs. If, as described above,
we're called in a different sched group than the original one however, this
can cause double registrations.

Fixed here by keeping a reference to creating scheduling group and using this, not
active one, when/if creating new metrics.

Closes #14294
2023-06-21 10:08:27 +03:00

136 lines
5.9 KiB
C++

/*
* Copyright (C) 2015-present ScyllaDB
*/
/*
* SPDX-License-Identifier: AGPL-3.0-or-later
*/
#include <seastar/core/thread.hh>
#include "test/lib/scylla_test_case.hh"
#include "query-result-writer.hh"
#include "test/lib/cql_test_env.hh"
#include "test/lib/mutation_source_test.hh"
#include "test/lib/result_set_assertions.hh"
#include "service/storage_proxy.hh"
#include "query_ranges_to_vnodes.hh"
#include "partition_slice_builder.hh"
#include "schema/schema_builder.hh"
// Returns random keys sorted in ring order.
// The schema must have a single bytes_type partition key column.
static std::vector<dht::ring_position> make_ring(schema_ptr s, int n_keys) {
std::vector<dht::ring_position> ring;
for (int i = 0; i < 10; ++i) {
auto pk = partition_key::from_single_value(*s, to_bytes(format("key{:d}", i)));
ring.emplace_back(dht::decorate_key(*s, pk));
}
std::sort(ring.begin(), ring.end(), dht::ring_position_less_comparator(*s));
return ring;
}
SEASTAR_TEST_CASE(test_get_restricted_ranges) {
return do_with_cql_env([](cql_test_env& e) {
return seastar::async([] {
auto s = schema_builder("ks", "cf")
.with_column("pk", bytes_type, column_kind::partition_key)
.with_column("v", bytes_type, column_kind::regular_column)
.build();
std::vector<dht::ring_position> ring = make_ring(s, 10);
auto check = [&s](locator::token_metadata_ptr tmptr, dht::partition_range input,
dht::partition_range_vector expected) {
query_ranges_to_vnodes_generator ranges_to_vnodes(locator::make_splitter(tmptr), s, {input});
auto actual = ranges_to_vnodes(1000);
if (!std::equal(actual.begin(), actual.end(), expected.begin(), [&s](auto&& r1, auto&& r2) {
return r1.equal(r2, dht::ring_position_comparator(*s));
})) {
BOOST_FAIL(format("Ranges differ, expected {} but got {}", expected, actual));
}
};
{
// Ring with minimum token
auto tmptr = locator::make_token_metadata_ptr(locator::token_metadata::config{});
tmptr->update_topology(gms::inet_address("10.0.0.1"), {"dc1", "rack1"});
tmptr->update_normal_tokens(std::unordered_set<dht::token>({dht::minimum_token()}), gms::inet_address("10.0.0.1")).get();
check(tmptr, dht::partition_range::make_singular(ring[0]), {
dht::partition_range::make_singular(ring[0])
});
check(tmptr, dht::partition_range({ring[2]}, {ring[3]}), {
dht::partition_range({ring[2]}, {ring[3]})
});
}
{
auto tmptr = locator::make_token_metadata_ptr(locator::token_metadata::config{});
tmptr->update_topology(gms::inet_address("10.0.0.1"), {"dc1", "rack1"});
tmptr->update_normal_tokens(std::unordered_set<dht::token>({ring[2].token()}), gms::inet_address("10.0.0.1")).get();
tmptr->update_topology(gms::inet_address("10.0.0.2"), {"dc1", "rack1"});
tmptr->update_normal_tokens(std::unordered_set<dht::token>({ring[5].token()}), gms::inet_address("10.0.0.2")).get();
check(tmptr, dht::partition_range::make_singular(ring[0]), {
dht::partition_range::make_singular(ring[0])
});
check(tmptr, dht::partition_range::make_singular(ring[2]), {
dht::partition_range::make_singular(ring[2])
});
check(tmptr, dht::partition_range({{dht::ring_position::ending_at(ring[2].token()), false}}, {ring[3]}), {
dht::partition_range({{dht::ring_position::ending_at(ring[2].token()), false}}, {ring[3]})
});
check(tmptr, dht::partition_range({ring[3]}, {ring[4]}), {
dht::partition_range({ring[3]}, {ring[4]})
});
check(tmptr, dht::partition_range({ring[2]}, {ring[3]}), {
dht::partition_range({ring[2]}, {dht::ring_position::ending_at(ring[2].token())}),
dht::partition_range({{dht::ring_position::ending_at(ring[2].token()), false}}, {ring[3]})
});
check(tmptr, dht::partition_range({{ring[2], false}}, {ring[3]}), {
dht::partition_range({{ring[2], false}}, {dht::ring_position::ending_at(ring[2].token())}),
dht::partition_range({{dht::ring_position::ending_at(ring[2].token()), false}}, {ring[3]})
});
}
});
});
}
SEASTAR_THREAD_TEST_CASE(test_split_stats) {
auto ep1 = gms::inet_address("127.0.0.1");
auto sg1 = create_scheduling_group("apa1", 100).get();
auto sg2 = create_scheduling_group("apa2", 100).get();
std::optional<service::storage_proxy_stats::split_stats> stats1, stats2;
// pretending to be abstract_write_response_handler type.
// created in various scheduling groups, in which they
// instantiate group-local split_stats.
with_scheduling_group(sg1, [&] {
stats1.emplace("tuta", "nils", "en nils", "nilsa", true);
}).get0();
with_scheduling_group(sg2, [&] {
stats2.emplace("tuta", "nils", "en nils", "nilsa", true);
}).get0();
// simulating the calling of storage_proxy::on_down, from gossip
// on node dropping out. If inside a write operation, we'll pick up
// write handlers and to "timeout_cb" on them, which in turn might
// call get_ep_stat, which evenually calls register_metrics for
// the DC written to.
// Point being is that either the above should not happen, or
// split_stats should be resilient to being called from different
// scheduling group.
stats1->register_metrics_for("DC1", ep1);
stats2->register_metrics_for("DC1", ep1);
}