Files
scylladb/db/system_distributed_keyspace.hh
Tomasz Grabiec 9daed59af9 Merge 'Tablet-aware restore' from Pavel Emelyanov
The mechanics of the restore is like this

- A /storage_service/tablets/restore API is called with (keyspace, table, endpoint, bucket, manifests) parameters
  - First, it populates the system_distributed.snapshot_sstables table with the data read from the manifests
  - Then it emplaces a bunch of tablet transitions (of a new "restore" kind), one for each tablet
- The topology coordinator handles the "restore" transition by calling a new RESTORE_TABLET RPC against all the current tablet replicas
- Each replica handles the RPC verb by
  - Reading the snapshot_sstables table
  - Filtering the read sstable infos against current node and tablet being handled
  - Downloading and attaching the filtered sstables

This PR includes system_distributed.snapshot_sstables table from @robertbindar and preparation work from @kreuzerkrieg that extracts raw sstables downloading and attaching from existing generic sstables loading code.

This is first step towards SCYLLADB-197 and lacks many things. In particular
- the API only works for single-DC cluster
- the caller needs to "lock" tablet boundaries with min/max tablet count
- not abortable
- no progress tracking
- sub-optimal (re-kicking API on restore will re-download everything again)
- not re-attacheable (if API node dies, restoration proceeds, but the caller cannot "wait" for it to complete via other node)
- nodes download sstables in maintenance/streaming sched gorup (should be moved to maintenance/backup)

Other follow-up items:
- have an actual swagger object specification for `backup_location`

Closes #28436
Closes #28657
Closes #28773

Closes scylladb/scylladb#28763

* github.com:scylladb/scylladb:
  test: Add test for backup vs migration race
  test: Restore resilience test
  sstables_loader: Fail tablet-restore task if not all sstables were downloaded
  sstables_loader: mark sstables as downloaded after attaching
  sstables_loader: return shared_sstable from attach_sstable
  db: add update_sstable_download_status method
  db: add downloaded column to snapshot_sstables
  db: extract snapshot_sstables TTL into class constant
  test: Add a test for tablet-aware restore
  tablets: Implement tablet-aware cluster-wide restore
  messaging: Add RESTORE_TABLET RPC verb
  sstables_loader: Add method to download and attach sstables for a tablet
  tablets: Add restore_config to tablet_transition_info
  sstables_loader: Add restore_tablets task skeleton
  test: Add rest_client helper to kick newly introduced API endpoint
  api: Add /storage_service/tablets/restore endpoint skeleton
  sstables_loader: Add keyspace and table arguments to manfiest loading helper
  sstables_loader_helpers: just reformat the code
  sstables_loader_helpers: generalize argument and variable names
  sstables_loader_helpers: generalize get_sstables_for_tablet
  sstables_loader_helpers: add token getters for tablet filtering
  sstables_loader_helpers: remove underscores from struct members
  sstables_loader: move download_sstable and get_sstables_for_tablet
  sstables_loader: extract single-tablet SST filtering
  sstables_loader: make download_sstable static
  sstables_loader: fix formating of the new `download_sstable` function
  sstables_loader: extract single SST download into a function
  sstables_loader: add shard_id to minimal_sst_info
  sstables_loader: add function for parsing backup manifests
  split utility functions for creating test data from database_test
  export make_storage_options_config from lib/test_services
  rjson: Add helpers for conversions to dht::token and sstable_id
  Add system_distributed_keyspace.snapshot_sstables
  add get_system_distributed_keyspace to cql_test_env
  code: Add system_distributed_keyspace dependency to sstables_loader
  storage_service: Export export handle_raft_rpc() helper
  storage_service: Export do_tablet_operation()
  storage_service: Split transit_tablet() into two
  tablets: Add braces around tablet_transition_kind::repair switch
2026-04-21 02:27:24 +02:00

162 lines
7.1 KiB
C++

/*
* Copyright (C) 2018-present ScyllaDB
*/
/*
* SPDX-License-Identifier: LicenseRef-ScyllaDB-Source-Available-1.1
*/
#pragma once
#include "schema/schema_fwd.hh"
#include "service/qos/qos_common.hh"
#include "utils/UUID.hh"
#include "utils/chunked_vector.hh"
#include "cdc/generation_id.hh"
#include "db/consistency_level_type.hh"
#include "locator/host_id.hh"
#include "dht/token.hh"
#include "sstables/types.hh"
#include <seastar/core/future.hh>
#include <seastar/core/sstring.hh>
#include <seastar/util/bool_class.hh>
#include <optional>
#include <unordered_map>
namespace cql3 {
class query_processor;
}
namespace cdc {
class stream_id;
class topology_description;
class streams_version;
} // namespace cdc
namespace service {
class storage_proxy;
class migration_manager;
}
namespace db {
using is_downloaded = bool_class<class is_downloaded_tag>;
struct snapshot_sstable_entry {
sstables::sstable_id sstable_id;
dht::token first_token;
dht::token last_token;
sstring toc_name;
sstring prefix;
is_downloaded downloaded{is_downloaded::no};
};
class system_distributed_keyspace {
public:
static constexpr auto NAME = "system_distributed";
static constexpr auto NAME_EVERYWHERE = "system_distributed_everywhere";
static constexpr auto VIEW_BUILD_STATUS = "view_build_status";
static constexpr auto SERVICE_LEVELS = "service_levels";
/* Nodes use this table to communicate new CDC stream generations to other nodes. */
static constexpr auto CDC_TOPOLOGY_DESCRIPTION = "cdc_generation_descriptions";
/* Nodes use this table to communicate new CDC stream generations to other nodes.
* Resides in system_distributed_everywhere. */
static constexpr auto CDC_GENERATIONS_V2 = "cdc_generation_descriptions_v2";
/* This table is used by CDC clients to learn about available CDC streams. */
static constexpr auto CDC_DESC_V2 = "cdc_streams_descriptions_v2";
/* Used by CDC clients to learn CDC generation timestamps. */
static constexpr auto CDC_TIMESTAMPS = "cdc_generation_timestamps";
/* Previous version of the "cdc_streams_descriptions_v2" table.
* We use it in the upgrade procedure to ensure that CDC generations appearing
* in the old table also appear in the new table, if necessary. */
static constexpr auto CDC_DESC_V1 = "cdc_streams_descriptions";
/* This table is used by the backup and restore code to store per-sstable metadata.
* The data the coordinator node puts in this table comes from the snapshot manifests. */
static constexpr auto SNAPSHOT_SSTABLES = "snapshot_sstables";
static constexpr uint64_t SNAPSHOT_SSTABLES_TTL_SECONDS = std::chrono::seconds(std::chrono::days(3)).count();
/* Information required to modify/query some system_distributed tables, passed from the caller. */
struct context {
/* How many different token owners (endpoints) are there in the token ring? */
size_t num_token_owners;
};
private:
cql3::query_processor& _qp;
service::migration_manager& _mm;
service::storage_proxy& _sp;
bool _started = false;
bool _forced_cdc_timestamps_schema_sync = false;
public:
static std::vector<schema_ptr> all_distributed_tables();
static std::vector<schema_ptr> all_everywhere_tables();
system_distributed_keyspace(cql3::query_processor&, service::migration_manager&, service::storage_proxy&);
future<> start();
future<> start_workload_prioritization();
future<> stop();
bool started() const { return _started; }
future<> insert_cdc_generation(utils::UUID, const cdc::topology_description&, context);
future<std::optional<cdc::topology_description>> read_cdc_generation(utils::UUID);
future<> create_cdc_desc(db_clock::time_point, const cdc::topology_description&, context);
future<bool> cdc_desc_exists(db_clock::time_point, context);
// Reads and builds generation map - a map from generation timestamps to vector of all stream ids for that generation.
// Generations with timestamp >= `not_older_than` are returned, plus the one just before it (the straddling generation).
// Returns empty map if there are no generations with timestamp >= `not_older_than`.
// NOTE: there's a sibling `read_cdc_for_tablets_versioned_streams`, that reads the same data for tables backed by tablets. The data returned is the same.
// NOTE: currently used only by alternator
future<std::map<db_clock::time_point, cdc::streams_version>> cdc_get_versioned_streams(db_clock::time_point not_older_than, context);
// Read current generation timestamp for the given table. Throws runtime_error (see `cql3::untyped_result_set::one()`) if table not found.
// NOTE: there's a sibling `read_cdc_for_tablets_current_generation_timestamp` in `system_keyspace`, that does the same for tables backed up by tablets.
// NOTE: currently used only by alternator
future<db_clock::time_point> cdc_current_generation_timestamp(context);
future<qos::service_levels_info> get_service_levels(qos::query_context ctx) const;
future<qos::service_levels_info> get_service_level(sstring service_level_name) const;
future<> set_service_level(sstring service_level_name, qos::service_level_options slo) const;
future<> drop_service_level(sstring service_level_name) const;
/* Inserts a single SSTable entry for a given snapshot, keyspace, table, datacenter,
* and rack. The row is written with the specified TTL (in seconds). Uses consistency
* level `EACH_QUORUM` by default.*/
future<> insert_snapshot_sstable(sstring snapshot_name, sstring ks, sstring table, sstring dc, sstring rack, sstables::sstable_id sstable_id, dht::token first_token, dht::token last_token, sstring toc_name, sstring prefix, is_downloaded downloaded, db::consistency_level cl = db::consistency_level::EACH_QUORUM);
/* Retrieves all SSTable entries for a given snapshot, keyspace, table, datacenter, and rack.
* If `start_token` and `end_token` are provided, only entries whose `first_token` is in the range [`start_token`, `end_token`] will be returned.
* Returns a vector of `snapshot_sstable_entry` structs containing `sstable_id`, `first_token`, `last_token`,
* `toc_name`, and `prefix`. Uses consistency level `LOCAL_QUORUM` by default. */
future<utils::chunked_vector<snapshot_sstable_entry>> get_snapshot_sstables(sstring snapshot_name, sstring ks, sstring table, sstring dc, sstring rack, db::consistency_level cl = db::consistency_level::LOCAL_QUORUM, std::optional<dht::token> start_token = std::nullopt, std::optional<dht::token> end_token = std::nullopt) const;
future<> update_sstable_download_status(sstring snapshot_name,
sstring ks,
sstring table,
sstring dc,
sstring rack,
sstables::sstable_id sstable_id,
dht::token start_token,
is_downloaded downloaded) const;
private:
future<> create_tables(std::vector<schema_ptr> tables);
};
}