scylladb/replica/distributed_loader.cc

/*
 * Copyright (C) 2018-present ScyllaDB
 */

/*
 * SPDX-License-Identifier: LicenseRef-ScyllaDB-Source-Available-1.0
 */

#include "db/view/view_building_worker.hh"
#include "sstables/shared_sstable.hh"
#include "utils/assert.hh"
#include <fmt/std.h>
#include <seastar/core/coroutine.hh>
#include <seastar/core/smp.hh>
#include <seastar/coroutine/maybe_yield.hh>
#include <seastar/coroutine/parallel_for_each.hh>
#include <seastar/util/closeable.hh>
#include "distributed_loader.hh"
#include "replica/database.hh"
#include "replica/global_table_ptr.hh"
#include "db/config.hh"
#include "db/extensions.hh"
#include "db/system_keyspace.hh"
#include "db/system_distributed_keyspace.hh"
#include "db/schema_tables.hh"
#include "compaction/compaction_manager.hh"
#include "compaction/task_manager_module.hh"
#include "sstables/sstables.hh"
#include "sstables/sstables_manager.hh"
#include "sstables/sstable_directory.hh"
#include "auth/common.hh"
#include "tracing/trace_keyspace_helper.hh"
#include "db/view/view_update_checks.hh"
#include <unordered_map>
#include "db/view/view_builder.hh"

extern logging::logger dblog;

static const std::unordered_set<std::string_view> system_keyspaces = {
                db::system_keyspace::NAME, db::schema_tables::NAME,
};

bool is_system_keyspace(std::string_view name) {
    return system_keyspaces.contains(name);
}

static const std::unordered_set<std::string_view> internal_keyspaces = {
        db::system_distributed_keyspace::NAME,
        db::system_distributed_keyspace::NAME_EVERYWHERE,
        db::system_keyspace::NAME,
        db::schema_tables::NAME,
        auth::meta::legacy::AUTH_KS,
        tracing::trace_keyspace_helper::KEYSPACE_NAME
};

bool is_internal_keyspace(std::string_view name) {
    return internal_keyspaces.contains(name);
}

namespace replica {

static io_error_handler error_handler_for_upload_dir() {
    return [] (std::exception_ptr eptr) {
        // do nothing about sstable exception and caller will just rethrow it.
    };
}

io_error_handler error_handler_gen_for_upload_dir(disk_error_signal_type& dummy) {
    return error_handler_for_upload_dir();
}

future<>
distributed_loader::process_sstable_dir(sharded<sstables::sstable_directory>& dir, sstables::sstable_directory::process_flags flags) {
    co_await dir.invoke_on(0, [flags] (sstables::sstable_directory& d) -> future<> {
        co_await d.prepare(flags);
    });

    co_await dir.invoke_on_all([&dir, flags] (sstables::sstable_directory& d) -> future<> {
        // Supposed to be called with the node either down or on behalf of maintenance tasks
        // like nodetool refresh
        co_await d.process_sstable_dir(flags);
        co_await d.move_foreign_sstables(dir);
    });

    co_await dir.invoke_on_all(&sstables::sstable_directory::commit_directory_changes);
}

future<>
distributed_loader::lock_table(global_table_ptr& table, sharded<sstables::sstable_directory>& dir) {
    return dir.invoke_on_all([&table] (sstables::sstable_directory& d) {
        d.store_phaser(table->write_in_progress());
        return make_ready_future<>();
    });
}

// Global resharding function. Done in two parts:
//  - The first part spreads the foreign_sstable_open_info across shards so that all of them are
//    resharding about the same amount of data
//  - The second part calls each shard's distributed object to reshard the SSTables they were
//    assigned.
future<>
distributed_loader::reshard(sharded<sstables::sstable_directory>& dir, sharded<replica::database>& db, sstring ks_name, sstring table_name, sstables::compaction_sstable_creator_fn creator, compaction::owned_ranges_ptr owned_ranges_ptr) {
    auto& compaction_module = db.local().get_compaction_manager().get_task_manager_module();
    auto task = co_await compaction_module.make_and_start_task<table_resharding_compaction_task_impl>({}, std::move(ks_name), std::move(table_name), dir, db, std::move(creator), std::move(owned_ranges_ptr));
    co_await task->done();
}

future<sstables::sstable::version_types>
highest_version_seen(sharded<sstables::sstable_directory>& dir, sstables::sstable_version_types system_version) {
    using version = sstables::sstable_version_types;
    return dir.map_reduce0(std::mem_fn(&sstables::sstable_directory::highest_version_seen), system_version, [] (version a, version b) {
        return std::max(a, b);
    });
}

future<>
distributed_loader::reshape(sharded<sstables::sstable_directory>& dir, sharded<replica::database>& db, sstables::reshape_mode mode,
        sstring ks_name, sstring table_name, sstables::compaction_sstable_creator_fn creator,
        std::function<bool (const sstables::shared_sstable&)> filter) {
    auto& compaction_module = db.local().get_compaction_manager().get_task_manager_module();
    auto task = co_await compaction_module.make_and_start_task<table_reshaping_compaction_task_impl>({}, std::move(ks_name), std::move(table_name), dir, db, mode, std::move(creator), std::move(filter));
    co_await task->done();
}

// Loads SSTables into the main directory (or staging) and returns how many were loaded
future<size_t>
distributed_loader::make_sstables_available(sstables::sstable_directory& dir, sharded<replica::database>& db,
        sharded<db::view::view_builder>& vb, sharded<db::view::view_building_worker>& vbw, db::view::sstable_destination_decision needs_view_update, sstring ks, sstring cf) {

    auto& table = db.local().find_column_family(ks, cf);
    auto new_sstables = std::vector<sstables::shared_sstable>();

    co_await dir.do_for_each_sstable([&table, needs_view_update, &new_sstables] (sstables::shared_sstable sst) -> future<> {
        auto gen = table.calculate_generation_for_new_table();
        dblog.trace("Loading {} into {}, new generation {}", sst->get_filename(), needs_view_update == db::view::sstable_destination_decision::normal_directory ? "base" : "staging", gen);
        co_await sst->pick_up_from_upload(needs_view_update == db::view::sstable_destination_decision::normal_directory ? sstables::sstable_state::normal : sstables::sstable_state::staging, gen);
        // When loading an imported sst, set level to 0 because it may overlap with existing ssts on higher levels.
        sst->set_sstable_level(0);
        new_sstables.push_back(std::move(sst));
    });

    // nothing loaded
    if (new_sstables.empty()) {
        co_return 0;
    }

    co_await table.add_sstables_and_update_cache(new_sstables).handle_exception([&table] (std::exception_ptr ep) {
        dblog.error("Failed to load SSTables for {}.{}: {}. Aborting.", table.schema()->ks_name(), table.schema()->cf_name(), ep);
        abort();
    });

    if (needs_view_update == db::view::sstable_destination_decision::staging_managed_by_vbc) {
        co_await vbw.local().register_staging_sstable_tasks(new_sstables, table.schema()->id());
    } else if (needs_view_update == db::view::sstable_destination_decision::staging_directly_to_generator) {
        co_await coroutine::parallel_for_each(new_sstables, [&vb, &table] (sstables::shared_sstable sst) -> future<> {
            if (sst->requires_view_building()) {
                co_await vb.local().register_staging_sstable(sst, table.shared_from_this());
            }
        });
    }

    co_return new_sstables.size();
}

future<>
distributed_loader::process_upload_dir(sharded<replica::database>& db, sharded<db::view::view_builder>& vb, sharded<db::view::view_building_worker>& vbw, sstring ks, sstring cf, bool skip_cleanup, bool skip_reshape) {
    const auto& rs = db.local().find_keyspace(ks).get_replication_strategy();
    if (rs.is_per_table()) {
        on_internal_error(dblog, "process_upload_dir is not supported with tablets");
    }

    return seastar::async([&db, &vb, &vbw, ks = std::move(ks), cf = std::move(cf), skip_cleanup, skip_reshape] {
        auto global_table = get_table_on_all_shards(db, ks, cf).get();

        sharded<sstables::sstable_directory> directory;
        directory.start(global_table.as_sharded_parameter(),
            sstables::sstable_state::upload, &error_handler_gen_for_upload_dir
        ).get();

        auto stop_directory = deferred_stop(directory);

        lock_table(global_table, directory).get();
        sstables::sstable_directory::process_flags flags {
            .need_mutate_level = true,
            .enable_dangerous_direct_import_of_cassandra_counters = db.local().get_config().enable_dangerous_direct_import_of_cassandra_counters(),
            .allow_loading_materialized_view = false,
        };
        process_sstable_dir(directory, flags).get();

        auto make_sstable = [&] (shard_id shard) {
            auto& sstm = global_table->get_sstables_manager();
            auto& gen = global_table->get_sstable_generation_generator();
            auto generation = gen();
            return sstm.make_sstable(global_table->schema(), global_table->get_storage_options(),
                                     generation, sstables::sstable_state::upload, sstm.get_preferred_sstable_version(),
                                     sstables::sstable_format_types::big, db_clock::now(), &error_handler_gen_for_upload_dir);
        };
        // Pass owned_ranges_ptr to reshard to piggy-back cleanup on the resharding compaction.
        // Note that needs_cleanup() is inaccurate and may return false positives,
        // maybe triggering resharding+cleanup unnecessarily for some sstables.
        // But this is resharding on refresh (sstable loading via upload dir),
        // which will usually require resharding anyway.
        //
        // FIXME: take multiple compaction groups into account
        // - segregate resharded tables into compaction groups
        // - split the keyspace local ranges per compaction_group as done in table::perform_cleanup_compaction
        //   so that cleanup can be considered per compaction group
        const auto& erm = db.local().find_keyspace(ks).get_static_effective_replication_map();
        auto owned_ranges_ptr = skip_cleanup ? lw_shared_ptr<dht::token_range_vector>(nullptr) : compaction::make_owned_ranges_ptr(db.local().get_keyspace_local_ranges(erm).get());
        reshard(directory, db, ks, cf, make_sstable, owned_ranges_ptr).get();
        if (!skip_reshape) {
            reshape(directory, db, sstables::reshape_mode::strict, ks, cf, make_sstable,
                    [] (const sstables::shared_sstable&) { return true; }).get();
        }

        // Move to staging directory to avoid clashes with future uploads. Unique generation number ensures no collisions.
        const auto use_view_update_path = db::view::check_needs_view_update_path(vb.local(), erm->get_token_metadata_ptr(), *global_table, streaming::stream_reason::repair).get();

        size_t loaded = directory.map_reduce0([&db, ks, cf, use_view_update_path, &vb, &vbw] (sstables::sstable_directory& dir) {
            return make_sstables_available(dir, db, vb, vbw, use_view_update_path, ks, cf);
        }, size_t(0), std::plus<size_t>()).get();

        dblog.info("Loaded {} SSTables", loaded);
    });
}

future<std::tuple<table_id, std::vector<std::vector<sstables::shared_sstable>>>>
distributed_loader::get_sstables_from_upload_dir(sharded<replica::database>& db, sstring ks, sstring cf, sstables::sstable_open_config cfg) {
    return get_sstables_from(db, ks, cf, cfg, [] (auto& global_table, auto& directory) {
        return directory.start(global_table.as_sharded_parameter(),
            sstables::sstable_state::upload, &error_handler_gen_for_upload_dir
        );
    });
}

future<std::tuple<table_id, std::vector<std::vector<sstables::shared_sstable>>>>
distributed_loader::get_sstables_from_object_store(sharded<replica::database>& db, sstring ks, sstring cf, std::vector<sstring> sstables, sstring endpoint, sstring bucket, sstring prefix, sstables::sstable_open_config cfg, std::function<seastar::abort_source*()> get_abort_src) {
    return get_sstables_from(db, ks, cf, cfg, [bucket, endpoint, prefix, sstables=std::move(sstables), &get_abort_src] (auto& global_table, auto& directory) {
        return directory.start(global_table.as_sharded_parameter(),
            sharded_parameter([bucket, endpoint, prefix, &get_abort_src] {
                data_dictionary::storage_options opts;
                seastar::abort_source* as = get_abort_src ? get_abort_src() : nullptr;
                opts.value = data_dictionary::storage_options::s3{bucket, endpoint, prefix, as};
                return make_lw_shared<const data_dictionary::storage_options>(std::move(opts));
            }),
            sstables,
            &error_handler_gen_for_upload_dir);
    });
}

future<std::tuple<table_id, std::vector<std::vector<sstables::shared_sstable>>>>
distributed_loader::get_sstables_from(sharded<replica::database>& db, sstring ks, sstring cf, sstables::sstable_open_config cfg,
        noncopyable_function<future<>(global_table_ptr&, sharded<sstables::sstable_directory>&)> start_dir) {
    return seastar::async([&db, ks = std::move(ks), cf = std::move(cf), start_dir = std::move(start_dir), cfg] {
        auto global_table = get_table_on_all_shards(db, ks, cf).get();
        auto table_id = global_table->schema()->id();

        sharded<sstables::sstable_directory> directory;
        start_dir(global_table, directory).get();
        auto stop = deferred_stop(directory);

        std::vector<std::vector<sstables::shared_sstable>> sstables_on_shards(smp::count);
        lock_table(global_table, directory).get();
        sstables::sstable_directory::process_flags flags {
            .need_mutate_level = true,
            .enable_dangerous_direct_import_of_cassandra_counters = db.local().get_config().enable_dangerous_direct_import_of_cassandra_counters(),
            .allow_loading_materialized_view = false,
            .sort_sstables_according_to_owner = false,
            .sstable_open_config = cfg,
        };
        process_sstable_dir(directory, flags).get();
        directory.invoke_on_all([&sstables_on_shards] (sstables::sstable_directory& d) mutable {
            sstables_on_shards[this_shard_id()] = d.get_unsorted_sstables();
        }).get();

        return std::make_tuple(table_id, std::move(sstables_on_shards));
    });
}

class table_populator {
    sharded<replica::database>& _db;
    sstring _ks;
    sstring _cf;
    global_table_ptr& _global_table;
    std::vector<lw_shared_ptr<sharded<sstables::sstable_directory>>> _sstable_directories;
    sstables::sstable_version_types _version_for_reshaping = sstables::oldest_writable_sstable_format;

public:
    table_populator(global_table_ptr& ptr, sharded<replica::database>& db, sstring ks, sstring cf)
        : _db(db)
        , _ks(std::move(ks))
        , _cf(std::move(cf))
        , _global_table(ptr)
    {
    }

    ~table_populator() {
        // All directories must have been stopped
        // using table_populator::stop()
        SCYLLA_ASSERT(_sstable_directories.empty());
    }

    future<> start();
    future<> stop();

private:
    using allow_offstrategy_compaction = bool_class<struct allow_offstrategy_compaction_tag>;

    future<> collect_subdirs();
    future<> collect_subdirs(const data_dictionary::storage_options::local&, sstables::sstable_state state);
    future<> collect_subdirs(const data_dictionary::storage_options::s3&, sstables::sstable_state state);
    future<> process_subdir(sharded<sstables::sstable_directory>&);
    future<> populate_subdir(sharded<sstables::sstable_directory>&);
};

future<> table_populator::start() {
    SCYLLA_ASSERT(this_shard_id() == 0);

    co_await collect_subdirs();

    for (auto dir : _sstable_directories) {
        co_await process_subdir(*dir);
    }

    co_await smp::invoke_on_all([this] {
        return _global_table->disable_auto_compaction();
    });

    for (auto dir : _sstable_directories) {
        co_await populate_subdir(*dir);
    }
}

future<> table_populator::stop() {
    while (!_sstable_directories.empty()) {
        co_await _sstable_directories.back()->stop();
        _sstable_directories.pop_back();
    }
}

future<> table_populator::collect_subdirs(const data_dictionary::storage_options::local& so, sstables::sstable_state state) {
    auto datadirs = _global_table->get_sstables_manager().get_local_directories(so);
    co_await coroutine::parallel_for_each(datadirs, [&] (const std::filesystem::path& datadir) -> future<> {
        auto dptr = make_lw_shared<sharded<sstables::sstable_directory>>();
        co_await dptr->start(_global_table.as_sharded_parameter(), state,
                        sharded_parameter([datadir] {
                            auto opts = data_dictionary::make_local_options(datadir);
                            return make_lw_shared<const data_dictionary::storage_options>(std::move(opts));
                        }), default_io_error_handler_gen());
        _sstable_directories.push_back(std::move(dptr));
    });
}

future<> table_populator::collect_subdirs(const data_dictionary::storage_options::s3& so, sstables::sstable_state state) {
    auto dptr = make_lw_shared<sharded<sstables::sstable_directory>>();
    co_await dptr->start(_global_table.as_sharded_parameter(), state, default_io_error_handler_gen());
    _sstable_directories.push_back(std::move(dptr));
}

future<> table_populator::collect_subdirs() {
    // The table base directory (with sstable_state::normal) must be
    // loaded and processed first as it now may contain the shared
    // pending_delete_dir, possibly referring to sstables in sub-directories.
    for (auto state : { sstables::sstable_state::normal, sstables::sstable_state::staging, sstables::sstable_state::quarantine }) {
        co_await std::visit([this, state] (const auto& so) -> future<> { co_await collect_subdirs(so, state); }, _global_table->get_storage_options().value);
    }
    // directory must be stopped using table_populator::stop below
}

future<> table_populator::process_subdir(sharded<sstables::sstable_directory>& directory) {
    co_await distributed_loader::lock_table(_global_table, directory);

    sstables::sstable_directory::process_flags flags {
        .throw_on_missing_toc = true,
        .enable_dangerous_direct_import_of_cassandra_counters = _db.local().get_config().enable_dangerous_direct_import_of_cassandra_counters(),
        .allow_loading_materialized_view = true,
        .garbage_collect = true,
    };
    co_await distributed_loader::process_sstable_dir(directory, flags);

    // If we are resharding system tables before we can read them, we will not
    // know which is the highest format we support: this information is itself stored
    // in the system tables. In that case we'll rely on what we find on disk: we'll
    // at least not downgrade any files. If we already know that we support a higher
    // format than the one we see then we use that.
    auto sst_version = co_await highest_version_seen(directory, sstables::oldest_writable_sstable_format);
    _version_for_reshaping = _global_table->get_sstables_manager().get_safe_sstable_version_for_rewrites(sst_version);
}

sstables::shared_sstable make_sstable(replica::table& table, sstables::sstable_state state, sstables::generation_type generation, sstables::sstable_version_types v) {
    return table.get_sstables_manager().make_sstable(table.schema(), table.get_storage_options(), generation, state, v, sstables::sstable_format_types::big);
}

future<> table_populator::populate_subdir(sharded<sstables::sstable_directory>& directory) {
    auto state = directory.local().state();
    dblog.debug("Populating {}/{}/{} state={}", _ks, _cf, _global_table->get_storage_options(), state);

    co_await distributed_loader::reshard(directory, _db, _ks, _cf, [this, state] (shard_id shard) mutable {
        auto gen = smp::submit_to(shard, [this] () {
            return _global_table->calculate_generation_for_new_table();
        }).get();

        return make_sstable(*_global_table, state, gen, _version_for_reshaping);
    });

    // The node is offline at this point so we are very lenient with what we consider
    // offstrategy.
    // SSTables created by repair may not conform to compaction strategy layout goal
    // because data segregation is only performed by compaction
    // Instead of reshaping them on boot, let's add them to maintenance set and allow
    // off-strategy compaction to reshape them. This will allow node to become online
    // ASAP. Given that SSTables with repair origin are disjoint, they can be efficiently
    // read from.
    auto eligible_for_reshape_on_boot = [] (const sstables::shared_sstable& sst) {
        return sst->get_origin() != sstables::repair_origin;
    };

    co_await distributed_loader::reshape(directory, _db, sstables::reshape_mode::relaxed, _ks, _cf, [this, state](shard_id shard) {
        auto gen = _global_table->calculate_generation_for_new_table();
        return make_sstable(*_global_table, state, gen, _version_for_reshaping);
    }, eligible_for_reshape_on_boot);

    auto do_allow_offstrategy_compaction = allow_offstrategy_compaction(state == sstables::sstable_state::normal);
    co_await directory.invoke_on_all([this, &eligible_for_reshape_on_boot, do_allow_offstrategy_compaction] (sstables::sstable_directory& dir) -> future<> {
        co_await dir.do_for_each_sstable([this, &eligible_for_reshape_on_boot, do_allow_offstrategy_compaction] (sstables::shared_sstable sst) {
            auto requires_offstrategy = sstables::offstrategy(do_allow_offstrategy_compaction && !eligible_for_reshape_on_boot(sst));
            return _global_table->add_sstable_and_update_cache(sst, requires_offstrategy);
        });
        if (do_allow_offstrategy_compaction) {
            _global_table->trigger_offstrategy_compaction();
        }
    });
}

future<> distributed_loader::populate_keyspace(sharded<replica::database>& db,
        sharded<db::system_keyspace>& sys_ks, keyspace& ks, sstring ks_name)
{
    dblog.info("Populating Keyspace {}", ks_name);

    co_await coroutine::parallel_for_each(ks.metadata()->cf_meta_data() | std::views::values, [&] (schema_ptr s) -> future<> {
        auto uuid = s->id();
        sstring cfname = s->cf_name();
        auto gtable = co_await get_table_on_all_shards(db, ks_name, cfname);
        auto& cf = *gtable;

        dblog.info("Keyspace {}: Reading CF {} id={} version={} storage={}", ks_name, cfname, uuid, s->version(), cf.get_storage_options());

        auto metadata = table_populator(gtable, db, ks_name, cfname);
        std::exception_ptr ex;

        try {
            co_await metadata.start();
        } catch (...) {
            std::exception_ptr eptr = std::current_exception();
            std::string msg =
                format("Exception while populating keyspace '{}' with column family '{}' from '{}': {}",
                        ks_name, cfname, cf.get_storage_options(), eptr);
            dblog.error("{}", msg);
            try {
                std::rethrow_exception(eptr);
            } catch (sstables::compaction_stopped_exception& e) {
                // swallow compaction stopped exception, to allow clean shutdown.
            } catch (...) {
                ex = std::make_exception_ptr(std::runtime_error(msg.c_str()));
            }
        }

        co_await metadata.stop();
        if (ex) {
            co_await coroutine::return_exception_ptr(std::move(ex));
        }

        // system tables are made writable through sys_ks::mark_writable
        if (!is_system_keyspace(ks_name)) {
            co_await smp::invoke_on_all([&] {
                auto s = gtable->schema();
                db.local().find_column_family(s).mark_ready_for_writes(db.local().commitlog_for(s));
            });
        }
    });

    // here we can be sure that the 'truncated' table is loaded,
    // now we load and initialize the truncation times for the tables
    {
        const auto truncation_times = co_await sys_ks.local().load_truncation_times();
        co_await smp::invoke_on_all([&] {
            db.local().get_tables_metadata().for_each_table([&](table_id id, lw_shared_ptr<table> table) {
                if (table->schema()->ks_name() != ks_name) {
                    return;
                }
                const auto it = truncation_times.find(id);
                const auto truncation_time = it == truncation_times.end()
                    ? db_clock::time_point::min()
                    : it->second;
                if (this_shard_id() == 0 && table->schema()->ks_name() == db::schema_tables::NAME &&
                    truncation_time != db_clock::time_point::min()) {
                    // replay_position stored in the truncation record may belong to
                    // the old (default) commitlog domain. It's not safe to interpret
                    // that replay position in the schema commitlog domain.
                    // Refuse to boot in this case. We assume no one truncated schema tables.
                    // We will hit this during rolling upgrade, in which case the user will
                    // roll back and let us know.
                    throw std::runtime_error(format("Schema table {}.{} has a truncation record. Booting is not safe.",
                        table->schema()->ks_name(), table->schema()->cf_name()));
                }
                table->set_truncation_time(truncation_time);
            });
        });
    }
}

future<> distributed_loader::init_system_keyspace(sharded<db::system_keyspace>& sys_ks, sharded<locator::effective_replication_map_factory>& erm_factory, sharded<replica::database>& db) {
    return seastar::async([&sys_ks, &erm_factory, &db] {
        sys_ks.invoke_on_all([&erm_factory, &db] (auto& sys_ks) {
            return sys_ks.make(erm_factory.local(), db.local());
        }).get();

        for (auto ksname : system_keyspaces) {
            auto& ks = db.local().get_keyspaces();
            auto i = ks.find(ksname);
            if (i != ks.end()) {
                distributed_loader::populate_keyspace(db, sys_ks, i->second, sstring(ksname)).get();
            }
        }
    });
}

future<> distributed_loader::init_non_system_keyspaces(sharded<replica::database>& db,
        sharded<service::storage_proxy>& proxy, sharded<db::system_keyspace>& sys_ks) {
    return seastar::async([&db, &proxy, &sys_ks] {
        db.invoke_on_all([&proxy, &sys_ks] (replica::database& db) {
            return db.parse_system_tables(proxy, sys_ks);
        }).get();

        const auto& cfg = db.local().get_config();

        for (bool prio_only : { true, false}) {
            std::vector<future<>> futures;

            for (auto& ks : db.local().get_keyspaces()) {
                auto& ks_name = ks.first;

                if (is_system_keyspace(ks_name)) {
                    continue;
                }

                /**
                 * Must process in two phases: Prio and non-prio.
                 * This looks like it is not needed. And it is not
                 * in open-source version. But essential for enterprise.
                 * Do _not_ remove or refactor away.
                 */
                if (prio_only != cfg.extensions().is_extension_internal_keyspace(ks_name)) {
                    continue;
                }

                futures.emplace_back(distributed_loader::populate_keyspace(db, sys_ks, ks.second, ks_name));
            }

            when_all_succeed(futures.begin(), futures.end()).discard_result().get();
        }
    });
}

}