The intranode shard balancing loop only stopped when the most-loaded
and least-loaded shard were the same (src == dst), meaning it would
keep issuing migrations until the load difference reached exactly 0.
This caused unnecessary migrations for negligible imbalances.
Apply the same is_balanced() threshold check that is already used for
inter-node balancing, so that intranode migrations stop when the
relative load difference between shards is within the configured
size_based_balance_threshold (default 1%).
(cherry picked from commit aaead10e5d)
- Handle dropped tables gracefully in the tablet load balancer's `get_schema_and_rs()` instead of aborting with `on_internal_error`
- The load balancer operates on a token metadata snapshot but accesses the live schema for table lookups. A DROP TABLE applied by another fiber between coroutine yield points can remove a table from the live schema while it still exists in the snapshot, causing an abort.
`get_schema_and_rs()` now returns `std::optional` and logs a warning in debug log level instead of aborting when a table is missing. All callers skip dropped tables:
- `make_sizing_plan`: skips to next table
- `make_resize_plan`: skips to next table (merge suppression is moot)
- `check_constraints`: returns `skip_info{}` with empty viable targets
- `get_rs`: returns `nullptr`, checked by `check_constraints`
The call chain is: `make_plan` → `make_internode_plan` → `check_constraints` → `get_rs` → `get_schema_and_rs`. The `make_internode_plan` coroutine has multiple `co_await` yield points (`maybe_yield`, `pick_candidate`) between building the candidate tablet list and checking replication constraints. A DROP TABLE schema mutation applied during any of these yields removes the table from `_db.get_tables_metadata()` while the candidate list still references it.
Added `test_load_balancing_with_dropped_table` which simulates the race by capturing a token metadata snapshot, dropping the table, then calling `balance_tablets` with the stale snapshot.
Fixes: SCYLLADB-1905
This fix needs to be backported to versions: 2025.4, 2026.1
- (cherry picked from commit 4987204f71)
- (cherry picked from commit 6b3e18c4a9)
Parent PR: #29585Closesscylladb/scylladb#29818
* github.com:scylladb/scylladb:
test: verify load balancer handles dropped tables gracefully
tablet_allocator: handle dropped tables gracefully in get_schema_and_rs
When force_capacity_based_balancing is enabled, the tablet allocator
balances by node and shard capacity rather than by tablet sizes.
When the data needed for load balancing is incomplete, the balancer
fails and waits until load_stats is available and correct for all the
nodes. An exception to this is when a node is being drained and
excluded: it is unreachable, and will not return. In this case
the balancer has to do its best and ignore the missing data.
This patch fixes a bug where forcing capacity based balancing made the
balancer not ignore missing data in these cases, and instead abort the
balancing.
(cherry picked from commit 906d2b817e)
Fixes: SCYLLADB-1936
In case we abort a decommission operation, the snapshot/backup
mechanism need to remain open.
This change moves it to after raft_decommission.
In the case of a cluster snapshot, our nodes ownership
or not of tables will be serialized by raft anyway, so
should remain consistent. In that case we at worst coordinate
from a node in "leave" status
In the case of a local snapshot, ownership matters less,
only sstables on disk, which should not change.
In the case of backup, this operates on a snapshot, state of which
is not affected.
Adds an injection point for testing.
v2:
- Added injection point to ensure test can abort decommission
Closesscylladb/scylladb#29667
(cherry picked from commit 2cc1a2c406)
Closesscylladb/scylladb#29848
The load balancer's get_schema_and_rs() would trigger on_internal_error when
a table present in the token metadata snapshot had been concurrently dropped
from the live schema. This race is possible because the balancer coroutine
yields between building the candidate list and checking replication
constraints, allowing a DROP TABLE schema mutation to be applied by another
fiber in the meantime.
Change get_schema_and_rs() to return {nullptr, nullptr} for dropped tables
instead of aborting. Update all callers to skip dropped tables:
- make_sizing_plan: continue to next table
- make_resize_plan: continue to next table (merge suppression is moot)
- check_constraints: return skip_info with empty viable targets
- get_rs: return nullptr, checked by check_constraints
(cherry picked from commit 4987204f71)
Commit 2b7aa32 (topology_coordinator: Refresh load stats after
table is created or altered) registered topology_coordinator as a
schema change listener and added on_create_column_family which
fire-and-forgets _tablet_load_stats_refresh.trigger(). The
triggered task runs on the gossip scheduling group via
with_scheduling_group and accesses the topology_coordinator via
'this'.
stop() unregisters the listener but does not wait for any
in-flight refresh task. If a notification fires between
_tablet_load_stats_refresh.join() in run() and unregister_listener
in stop(), the scheduled task can outlive the topology_coordinator
and access freed memory after run_topology_coordinator's coroutine
frame is destroyed.
Wait for the refresh to complete in stop() after unregistering the
listener, ensuring no task can fire after destruction.
Fixes SCYLLADB-1728
Backport to 2026.1 and 2026.2, because the issue was introduced in 2b7aa32Closesscylladb/scylladb#29653
* https://github.com/scylladb/scylladb:
test: tablet_stats: reproduce shutdown refresh race
topology_coordinator: join tablet load stats refresh in stop()
(cherry picked from commit d9dd3bfe53)
Closesscylladb/scylladb#29686
Add more logging to barrier and drain rpc to try and pinpoint https://github.com/scylladb/scylladb/issues/26281
Bakport since we want to have it if it happens in the field.
Fixes: SCYLLADB-1836
Refs: #26281Closesscylladb/scylladb#29735
* https://github.com/scylladb/scylladb:
session, raft_topology: add periodic warnings for hung drain and stale version waits
session: add info-level logging to drain_closing_sessions
raft_topology: log sub-step progress in local_topology_barrier
raft_topology: log read_barrier progress in topology cmd handler
(cherry picked from commit b69d00b0a7)
Closesscylladb/scylladb#29763
If start_server_for_group0() successfully registers a server in
_raft_gr._servers but a subsequent step (e.g. enable_in_memory_state_machine())
throws, the server is never destroyed because abort_and_drain()/destroy()
check std::get_if<raft::group_id>(&_group0) which was only set after the
entire with_scheduling_group block completed.
Move _group0.emplace<raft::group_id>() inside the lambda, immediately after
start_server_for_group() succeeds, so that cleanup paths can always find
and destroy the registered server.
This fixes the assertion:
"raft_group_registry - stop(): server for group ... is not destroyed"
which manifests during shutdown after an upgrade where topology_state_load()
fails due to netw::unknown_address.
Backport: Yes, to 2026.1, 2026.2, as it causes a crash on upgrades
Refs: SCYLLADB-1217
Refs: CUSTOMER-340
Refs: CUSTOMER-335
Fixes: SCYLLADB-1809
Signed-off-by: Yaniv Kaul <yaniv.kaul@scylladb.com>
AI-assisted: Yes, Opencode/Opus 4.6
Closesscylladb/scylladb#29702
(cherry picked from commit 6179406467)
Closesscylladb/scylladb#29742
In do_execute_cql_with_timeout(), when the prepared statement was not found in the cache, we called qp.prepare() and stored the returned result_message::prepared in a local variable scoped to the 'if' block. We then extracted ps_ptr (a checked_weak_ptr to the prepared statement) from the message, let the message go out of scope at the end of the 'if', and used ps_ptr after a co_await on st->execute().
Since 3ac4e258e8 ("transport/messages: hold pinned prepared entry in PREPARE result"), result_message::prepared owns a strong pinned reference to the prepared cache entry. While qp.prepare() runs it also holds its own pin on the entry, so on return the entry has at least the pin owned by the returned message. As long as that message is alive, the cache entry cannot be purged and the weak handle inside ps_ptr remains promotable.
The lifetime gap manifested only in debug builds. qp.prepare() returns a ready future on the cache-miss path, so in release builds the co_await resumes synchronously: control flows from the assignment of ps_ptr straight into st->execute() with no opportunity for any other task (in particular, prepared cache invalidation triggered by a concurrent schema change) to run in between. Debug builds, however, force a reactor preemption point on every co_await even when the awaited future is ready. With prepared_msg already destroyed at the end of the 'if' block, the only remaining handle on the cache entry was the weak ps_ptr, and the preemption gave a concurrent cache purge
- triggered, for example, by Raft schema changes received during a node restart - the chance to drop the entry. The subsequent execute() then failed when promoting the weak pointer with
checked_ptr_is_null_exception.
The exception propagated out of the Paxos prepare path as a generic std::exception with no type information in the log, surfacing on the coordinator as:
WriteFailure: Failed to prepare ballot ... Replica errors:
host_id ... -> seastar::rpc::remote_verb_error (std::exception)
Hoist the result_message::prepared into the outer scope so the pinned cache entry stays alive across co_await st->execute(...), closing the window in which a concurrent cache purge could invalidate the weak handle.
Fixes SCYLLADB-1173
backport: the patch is simple, we can backport it to all versions with "LWT over tablets" feature. Note that the problem is only in test runs in debug configuration, production is not affected.
Closesscylladb/scylladb#29675
* https://github.com/scylladb/scylladb:
table_helper: retry insert prepare on concurrent cache invalidation
paxos_state: keep prepared message alive across statement execution
(cherry picked from commit 15f35577ed)
Closesscylladb/scylladb#29701
get_ownership() and abort_topology_request() forward work to shard 0
via container().invoke_on(0, ...) but the lambda captured 'this' and
accessed members through it instead of through the shard-0 'ss'
parameter. This means the lambda used the caller-shard's instance,
defeating the purpose of the forwarding.
Use the 'ss' parameter consistently so the operations run against the
correct shard-0 state.
Hold _async_gate in all REST-facing async operations so that stop()
drains in-flight requests before teardown, preventing use-after-free
crashes when REST calls race with shutdown.
A centralized gated() wrapper in set_storage_service (api/storage_service.cc)
automatically holds the gate for every REST handler registered there,
so new handlers get shutdown-safety by default.
run_with_api_lock_internal and run_with_no_api_lock hold _async_gate on
shard 0 as well, because REST requests arriving on any shard are forwarded
there for execution.
Methods that previously self-forwarded to shard 0 (mark_excluded,
prepare_for_tablets_migration, set_node_intended_storage_mode,
get_tablets_migration_status, finalize_tablets_migration) now assert
this_shard_id() == 0. Their REST handlers call them via
run_with_no_api_lock, which performs the shard-0 hop and gate hold
centrally.
Fixes: SCYLLADB-1415
Add hold_async_gate() public accessor for use by the REST registration
layer in a followup commit.
Convert run_with_no_api_lock to a coroutine so a followup commit can
hold the async gate across the entire forwarded operation.
No functional changes.
With this change, you can add or remove a DC(s) in a single ALTER KEYSPACE statement. It requires the keyspace to use rack list replication factor.
In existing approach, during RF change all tablet replicas are rebuilt at once. This isn't the case now. In global_topology_request::keyspace_rf_change the request is added to a ongoing_rf_changes - a new column in system.topology table. In a new column in system_schema.keyspaces - next_replication - we keep the target RF.
In make_rf_change_plan, load balancer schedules necessary migrations, considering the load of nodes and other pending tablet transitions. Requests from ongoing_rf_changes are processed concurrently, independently from one another. In each request racks are processed concurrently. No tablet replica will be removed until all required replicas are added. While adding replicas to each rack we always start with base tables and won't proceed with views until they are done (while removing - the other way around). The intermediary steps aren't reflected in schema. When the Rf change is finished:
- in system_schema.keyspaces:
- next_replication is cleared;
- new keyspace properties are saved;
- request is removed from ongoing_rf_changes;
- the request is marked as done in system.topology_requests.
Until the request is done, DESCRIBE KEYSPACE shows the replication_v2.
If a request hasn't started to remove replicas, it can be aborted using task manager. system.topology_requests::error is set (but the request isn't marked as done) and next_replication = replication_v2. This will be interpreted by load balancer, that will start the rollback of the request. After the rollback is done, we set the relevant system.topology_requests entry as done (failed), clear the request id from system.topology::ongoing_rf_changes, and remove next_replication.
Fixes: SCYLLADB-567.
No backport needed; new feature.
Closesscylladb/scylladb#24421
* github.com:scylladb/scylladb:
service: fix indentation
docs: update documentation
test: test multi RF changes
service: tasks: allow aborting ongoing RF changes
cql3: allow changing RF by more than one when adding or removing a DC
service: handle multi_rf_change
service: implement make_rf_change_plan
service: add keyspace_rf_change_plan to migration_plan
service: extend tablet_migration_info to handle rebuilds
service: split update_node_load_on_migration
service: rearrange keyspace_rf_change handler
db: add columns to system_schema.keyspaces
db: service: add ongoing_rf_changes to system.topology
gms: add keyspace_multi_rf_change feature
When DROP TABLE races with an in-flight DML on a strongly-consistent
table, the node aborts in `groups_manager::acquire_server()` because the
raft group has already been erased from `_raft_groups`.
A concurrent `DROP TABLE` may have already removed the table from database
registries and erased the raft group via `schedule_raft_group_deletion`.
The `schema.table()` in `create_operation_ctx()` might not fail though
because someone might be holding `lw_shared_ptr<table>`, so that the
table is dropped but the table object is still alive.
Fix by accepting table_id in acquire_server and checking that the table
still exists in the database via `find_column_family` before looking up
the raft group. If the table has been dropped, find_column_family
throws no_such_column_family instead of the node aborting via
on_internal_error. When the table does exist, acquire_server proceeds
to acquire state.gate; schedule_raft_group_deletion co_awaits
gate::close, so it will wait for the DML operation to complete before
erasing the group.
backport: not needed (not released feature)
Fixes SCYLLADB-1450
Closesscylladb/scylladb#29430
* github.com:scylladb/scylladb:
strong_consistency: fix crash when DROP TABLE races with in-flight DML
test: add regression test for DROP TABLE racing with in-flight DML
The statement_restrictions code is responsible for analyzing the WHERE
clause, deciding on the query plan (which index to use), and extracting
the partition and clustering keys to use for the index.
Currently, it suffers from repetition in making its decisions: there are 15
calls to expr::visit in statement_restrictions.cc, and 14 find_binop calls. This
reduces to 2 visits (one nested in the other) and 6 find_binop calls. The analysis
of binary operators is done once, then reused.
The key data structure introduced is the predicate. While an expression
takes inputs from the row evaluated, constants, and bind variables, and
produces a boolean result, predicates ask which values for a column (or
a number of columns) are needed to satisfy (part of) the WHERE clause.
The WHERE clause is then expressed as a conjunction of such predicates.
The analyzer uses the predicates to select the index, then uses the predicates
to compute the partition and clustering keys.
The refactoring is composed of these parts (but patches from different parts
are interspersed):
1. an exhaustive regression test is added as the first commit, to ensure behavior doesn't change
2. move computation from query time to prepare time
3. introduce, gradually enrich, and use predicates to implement the statement_restrictions API
Major refactoring, and no bugs fixed, so definitely not backporting.
Closesscylladb/scylladb#29114
* github.com:scylladb/scylladb:
cql3: statement_restrictions: replace has_eq_restriction_on_column with precomputed set
cql3: statement_restrictions: replace multi_column_range_accumulator_builder with direct predicate iteration
cql3: statement_restrictions: use predicate fields in build_get_clustering_bounds_fn
cql3: statement_restrictions: remove extract_single_column_restrictions_for_column
cql3: statement_restrictions: use predicate vectors in prepare_indexed_local
cql3: statement_restrictions: use predicate vector size for clustering prefix length
cql3: statement_restrictions: replace do_find_idx and is_supported_by with predicate-based versions
cql3: statement_restrictions: remove expression-based has_supporting_index and index_supports_some_column
cql3: statement_restrictions: replace multi-column and PK index support checks with predicate-based versions
cql3: statement_restrictions: add predicate-based index support checking
cql3: statement_restrictions: use pre-built single-column maps for index support checks
cql3: statement_restrictions: build clustering-prefix restrictions incrementally
cql3: statement_restrictions: build partition-range restrictions incrementally
cql3: statement_restrictions: build clustering-key single-column restrictions map incrementally
cql3: statement_restrictions: build partition-key single-column restrictions map incrementally
cql3: statement_restrictions: build non-primary-key single-column restrictions map incrementally
cql3: statement_restrictions: use tracked has_mc_clustering for _has_multi_column
cql3: statement_restrictions: track has-token state incrementally
cql3: statement_restrictions: track partition-key-empty state incrementally
cql3: statement_restrictions: track first multi-column predicate incrementally
cql3: statement_restrictions: track last clustering column incrementally
cql3: statement_restrictions: track clustering-has-slice incrementally
cql3: statement_restrictions: track has-multi-column-clustering incrementally
cql3: statement_restrictions: track clustering-empty state incrementally
cql3: statement_restrictions: replace restr bridge variable with pred.filter
cql3: statement_restrictions: convert single-column branch to use predicate properties
cql3: statement_restrictions: convert multi-column branch to use predicate properties
cql3: statement_restrictions: convert constructor loop to iterate over predicates
cql3: statement_restrictions: annotate predicates with operator properties
cql3: statement_restrictions: annotate predicates with is_not_null and is_multi_column
cql3: statement_restrictions: complete preparation early
cql3: statement_restrictions: convert expressions to predicates without being directed at a specific column
cql3: statement_restrictions: refine possible_lhs_values() function_call processing
cql3: statement_restrictions: return nullptr for function solver if not token
cql3: statement_restrictions: refine possible_lhs_values() subscript solving
cql3: statement_restrictions: return nullptr from possible_lhs_values instead of on_internal_error
cql3: statement_restrictions: convert possible_lhs_values into a solver
cql3: statement_restrictions: split _where to boolean factors in preparation for predicates conversion
cql3: statement_restrictions: refactor IS NOT NULL processing
cql3: statement_restrictions: fold add_single_column_nonprimary_key_restriction() into its caller
cql3: statement_restrictions: fold add_single_column_clustering_key_restriction() into its caller
cql3: statement_restrictions: fold add_single_column_partition_key_restriction() into its caller
cql3: statement_restrictions: fold add_token_partition_key_restriction() into its caller
cql3: statement_restrictions: fold add_multi_column_clustering_key_restriction() into its caller
cql3: statement_restrictions: avoid early return in add_multi_column_clustering_key_restrictions
cql3: statement_restrictions: fold add_is_not_restriction() into its caller
cql3: statement_restrictions: fold add_restriction() into its caller
cql3: statement_restrictions: remove possible_partition_token_values()
cql3: statement_restrictions: remove possible_column_values
cql3: statement_restrictions: pass schema to possible_column_values()
cql3: statement_restrictions: remove fallback path in solve()
cql3: statement_restrictions: reorder possible_lhs_column parameters
cql3: statement_restrictions: prepare solver for multi-column restrictions
cql3: statement_restrictions: add solver for token restriction on index
cql3: statement_restrictions: pre-analyze column in value_for()
cql3: statement_restrictions: don't handle boolean constants in multi_column_range_accumulator_builder
cql3: statement_restrictions: split range_from_raw_bounds into prepare phase and query phase
cql3: statement_restrictions: adjust signature of range_from_raw_bounds
cql3: statement_restrictions: split multi_column_range_accumulator into prepare-time and query-time phases
cql3: statement_restrictions: make get_multi_column_clustering_bounds a builder
cql3: statement_restrictions: multi-key clustering restrictions one layer deeper
cql3: statement_restrictions: push multi-column post-processing into get_multi_column_clustering_bounds()
cql3: statement_restrictions: pre-analyze single-column clustering key restrictions
cql3: statement_restrictions: wrap value_for_index_partition_key()
cql3: statement_restrictions: hide value_for()
cql3: statement_restrictions: push down clustering prefix wrapper one level
cql3: statement_restrictions: wrap functions that return clustering ranges
cql3: statement_restrictions: do not pass view schema back and forth
cql3: statement_restrictions: pre-analyze token range restrictions
cql3: statement_restrictions: pre-analyze partition key columns
cql3: statement_restrictions: do not collect subscripted partition key columns
cql3: statement_restrictions: split _partition_range_restrictions into three cases
cql3: statement_restrictions: move value_list, value_set to header file
cql3: statement_restrictions: wrap get_partition_key_ranges
cql3: statement_restrictions: prepare statement_restrictions for capturing `this`
test: statement_restrictions: add index_selection regression test
Prevent copying/moving, that can change the address, and instead enforce
using shared_ptr. Most of the code is already using shared_ptr, so the
changes aren't very large.
To forbid non-shared_ptr construction, the constructors are annotated
with a private_tag tag class.
DynamoDB Streams API can only convey a single parent per stream shard.
Tablet merges produce two parents, making them incompatible with
Alternator Streams. This series blocks tablet merges when streams are
active on a tablet table.
For CreateTable, a freshly created table has no pending merges, so
streams are enabled immediately with tablet merges blocked.
For UpdateTable on an existing table, stream enablement is deferred:
the user's intent is stored via `enable_requested`, tablet merges are
blocked (new merge decisions are suppressed and any active merge
decision is revoked), and the topology coordinator finalizes enablement
once no in-flight merges remain.
The topology coordinator is woken promptly on error injection release
and tablet split completion, reducing finalization latency from ~60s
to seconds.
`test_parent_children_merge` is marked xfail (merges are now blocked),
and downward (merge) steps are removed from `test_parent_filtering` and
`test_get_records_with_alternating_tablets_count`.
Not addressed here: using a topology request to preempt long-running
operations like repair (tracked in SCYLLADB-1304).
Refs SCYLLADB-461
Closesscylladb/scylladb#29224
* github.com:scylladb/scylladb:
topology: Wake coordinator promptly for stream enablement lifecycle
test/cluster: Test deferred stream enablement on tablet tables
alternator/streams: Block tablet merges when Alternator Streams are enabled
The topology coordinator sleeps on a condition variable between
iterations. Several events relevant to Alternator stream enablement
did not wake it, causing delays of up to 60s (the periodic load
stats refresh interval) at each step:
1. Error injection release: when a test disables the
delay_cdc_stream_finalization injection, the coordinator was
not notified. Add an on_disable callback mechanism to the error
injection framework (register_on_disable / unregister_on_disable)
so subsystems can react when an injection is released. The
topology coordinator uses this to broadcast its event.
2. Tablet split completion: after all local storage groups for a
table finish splitting, split_ready_seq_number is set but the
coordinator only discovered this via the periodic stats refresh.
Add an on_tablet_split_ready callback to topology_state_machine
that the coordinator sets to trigger_load_stats_refresh(). The
split monitor in storage_service calls it when all compaction
groups are split-ready, giving the coordinator fresh stats
immediately so it can finalize the resize.
These changes reduce test_deferred_stream_enablement_on_tablets
from ~120s to ~13s and fix a production issue where Alternator
stream enablement could be delayed by up to 60s at each step of
the lifecycle (error injection release, split completion).
DynamoDB Streams API can only convey a single parent per stream shard.
Tablet merges produce 2 parents, which is incompatible. When streams
are requested on a tablet table, block tablet merges via
tablet_merge_blocked (the allocator suppresses new merge decisions and
revokes any active merge decision).
add_stream_options() sets tablet_merge_blocked=true alongside
enabled=true, so CreateTable needs no special handling — the flag
is inert on vnode tables and immediately effective on tablet tables.
For UpdateTable, CDC enablement is deferred: store the user's intent
via enable_requested, and let the topology coordinator finalize
enablement once no in-progress merges remain. A new helper,
defer_enabling_streams_block_tablet_merges(), amends the CDC options
to this deferred state.
Disabling streams clears all flags, immediately re-allowing merges.
The tablet allocator accesses the merge-blocked flag through a
schema::tablet_merges_forbidden() accessor rather than reaching into
CDC options directly.
Mark test_parent_children_merge as xfail and remove downward
(merge) steps from tablet_multipliers in test_parent_filtering and
test_get_records_with_alternating_tablets_count.
When a table is loaded on startup during a vnodes-to-tablets migration
(forward or rollback), the `table_populator` runs a resharding
compaction.
Set the migration virtual task as parent of the resharding task. This
enables users to easily find all node-local resharding tasks related to
a particular migration.
Make `migration_virtual_task::make_task_id()` public so that the
`distributed_loader` can compute the migration's task ID.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
Add a virtual task that exposes in-progress vnodes-to-tablets migrations
through the task manager API.
The task is synthesized from the current migration state, so completed
migrations are not shown. Progress is reported as the number of nodes
that currently use tablets: it increases on the forward path and
decreases on rollback. For simplicity, per-node storage modes are not
exposed in the task status; callers that need them should use the
migration status REST endpoint.
Unlike regular tasks that use time-based UUIDs, this task uses
deterministic named UUIDs derived from the keyspace names. This keeps
the implementation simple (no need to persist them) and gives each
keyspace a stable task ID. The downside is that the start time of each
task is unknown and repeated migrations of the same keyspace
(migration -> rollback -> new migration) cannot be distinguished.
Introduce a new task manager module to keep them separate from other
tasks.
Add support for `wait()`. While its practical value is debatable
(migration is a manual procedure, rolling restart will interrupt it), it
keeps the task consistent with the task manager interface.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
`storage_service::get_tablets_migration_status()` reads a group0 virtual
table, so it requires group0 to be initialized.
When invoked via the migration REST API, this condition is satisfied
since the API is only available after joining group0. However, once this
function is integrated into the task API later in this series, the
assumption will no longer hold, as the task API is exposed earlier in
the startup process.
Add a guard to detect this condition and return a clear error message.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
`storage_service::get_tablets_migration_status()` returns the
keyspace-level migration status, indicating whether migration has not
started, is in progress, or has completed, and for migrating keyspaces
also returns per-node migration statuses. Rename it to
`get_tablets_migration_status_with_node_details()` and introduce a new
`get_tablets_migration_status()` that returns only the keyspace-level
status.
This prepares the function for reuse in the next patches, which will add
a virtual task for vnodes-to-tablets migrations. Several task-manager
paths will only need the keyspace-level migration state, not per-node
information.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
Using a string was sufficient while this status was only exposed through
the REST API, but the next patches will also consume it internally.
Use an enum for the internal representation and convert it back to the
existing string values in the REST API.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
ScyllaDB uses estimated_histogram in many places.
We already have a more efficient alternative: estimated_histogram_with_max. It is both CPU- and
memory-efficient, and it can be exported as Prometheus native histograms.
Its main limitation (which also has benefits) is that the bucket layout is fixed at compile time, so
histograms with different configurations cannot be mixed.
The end goal is to replace all uses of estimated_histogram in the codebase.
That migration requires a few small API adjustments, so it is done in steps.
This PR replaces estimated_histogram for CAS contention.
The PR includes a patch that adds functionality to the base approx_exponential_histogram, which will be used by the API.
The specific histograms are defined in a single place and cover the range 1-100; this makes future changes easy.
**New feature, no need to backport**
Closesscylladb/scylladb#29017
* github.com:scylladb/scylladb:
storage_proxy: migrate CAS contention histograms to estimated_histogram_with_max
estimated_histogram.hh: Add bucket offset and count to approx_exponential_histogram
The PR serves two purposes.
First, it makes the flag usage be consistent across multiple ways to load sstables components. For example, the sstable::load_metadata() doesn't set it (like .load() does) thus potentially refusing to load "corrupted" components, as the flag assumes.
Second, it removes the fanout of db.get_config().ignore_component_digest_mismatch() over the code. This thing is called pretty much everywhere to initialize the sstable_open_config, while the option in question is "scylla state" parameter, not "sstable opening" one.
Code cleanup, not backporting
Closesscylladb/scylladb#29513
* github.com:scylladb/scylladb:
sstables: Remove ignore_component_digest_mismatch from sstable_open_config
sstables: Move ignore_component_digest_mismatch initialization to constructor
sstables: Add ignore_component_digest_mismatch to sstables_manager config
Allow aborting an ongoing RF change using task manager.
RF change can only be aborted if:
- it is currently paused (existing);
- it is a multi-RF change that still has replicas to be added.
In the second case, we set error for the request in system.topology_requests
and set next_replication to replication_v2. This makes load balancer
roll back the RF change.
Extend keyspace_rf_change handler to handle multi_rf_change.
multi_rf_change is allowed only if we add or remove DCs and
the keyspace uses rack list replication factor. The handler
adds the request id to topology::ongoing_rf_changes.
The request is further processed by load balancer.
In make_rf_change_plan, load balancer schedules necessary migrations,
considering the load of nodes and other pending tablet transitions.
Requests from ongoing_rf_changes are processed concurrently, independently
from one another. In each request racks are processed concurrently.
No tablet replica will be removed until all required replicas are added.
While adding replicas to each rack we always start with base tables
and won't proceed with views until they are done (while removing - the other
way around).
Node availability is checked at two levels for extending actions:
1) In prepare_per_rack_rf_change_plan: the entire RF change request is
aborted if any node in the target dc+rack is down, or if there are
no live (non-excluded) nodes at all. Shrinking is never aborted.
2) In make_rf_change_plan: extending is skipped for a given round if
any normal, non-excluded node in the target dc+rack is missing from
the balanced node set. Shrinking always proceeds regardless.
The resulting behavior per node state combination (extending only):
- all up -> proceed
- some excluded + some up -> proceed (excluded nodes are skipped)
- any down node -> abort
- all excluded (no live) -> abort
When the last step is finished:
- in system_schema.keyspaces:
- next_replication is cleared;
- new keyspace properties are saved (if request succeeded);
- request is removed from ongoing_rf_changes;
- the request is marked as done in system.topology_requests.
Add keyspace_rf_change_plan to migration_plan.
The keyspace_rf_change_plan consists of:
- completion - info about the request for which all migrations are done. Only one
request can be completed at the time, even if more have finished migrations
(the rest will be completed later). Based on it:
- next_replication is cleared;
- new keyspace properties are saved (only if succeeded);
- request is removed from ongoing_rf_changes;
- the request is marked as done in system.topology_requests.
- aborts - info about requests that cannot complete because the required
rf change is impossible (e.g. no available nodes in a required rack).
Multiple requests can be aborted in a single plan. Based on each:
- next_replication is set to current_replication (rolling back);
- the request is marked as aborted with an error in system.topology_requests.
The scheduled rebuilds will be kept in migration_plan::_migrations.
Based on that the canonical_mutations are generated.
Add update_topology_state_with_mixed_change and use it if any schema
changes are required, i.e. if plan contains keyspace_rf_change_plan::completion.
Split update_node_load_on_migration into decrease_node_load and
increase_node_load - in the following changes for rebuilds we will
need only one of those at the time.
In the following changes, keyspace_rf_change handler will also consider
a change of RF by more than one. Rearrange the handler, so that it
first chooses a kind of RF change and then creates relevant updates.
Do not wrap the code in schedule_migration function, as we no longer
need a quick return possibility.
Following changes, will allow adding or removing all keyspace
replicas in a DC with a single ALTER KEYSPACE. For such operations,
the tablet load balancer needs to schedule rebuilds. To track
which RF change requests require rebuilds, we maintain a vector
of RF changes along with their ongoing rebuild phases.
Add a new ongoing_rf_changes column to system.topology to keep track
of those requests.
In this commit no data is kept in the new column.
This PR removes the power-of-two token constraint from vnodes-to-tablets migrations, allowing clusters with randomly generated tokens to migrate without manual token reassignment.
Previously, migrations required vnode tokens to be a power of two and aligned. In practice, these conditions are not met with Scylla's default random token assignment, so the constraint is a blocker for real-world use. With the introduction of arbitrary tablet boundaries in PR #28459, the tablet layer can now support arbitrary tablet boundaries. This PR builds on that capability to allow arbitrary vnode tokens during migration.
When the highest vnode token does not coincide with the end of the token ring, the vnode wraps around, but tablets do not support that. This is handled by splitting it into two tablets: one covering the tail end of the ring and one covering the beginning.
Testing has been updated accordingly: existing cluster tests now use randomly generated tokens instead of precomputed power-of-two values, and a new Boost test validates the wrap-around tablet boundary logic.
Fixes SCYLLADB-724.
New feature, no backport is needed.
Closesscylladb/scylladb#29319
* github.com:scylladb/scylladb:
test: Use arbitrary tokens in vnodes->tablets migration tests
test: boost: Add test for wrap-around vnodes
storage_service: Support vnodes->tablets migrations w/ arbitrary tokens
storage_service: Hoist migration precondition
This will make debugging of stalled tablet transitions easier. We saw
several issues when topology state machine was blocked by active
tablet migrations, which was not obvious at first glance of the
logs. Now it will be east to tell if tablet transitions are blocking
progress and which transitions are stuck.
Closesscylladb/scylladb#28616
The ignore_component_digest_mismatch flag is now initialized at sstable construction
time from sstables_manager::config (which is populated from db::config at boot time).
Remove the flag from sstable_open_config struct and all call sites that were setting
it explicitly.
Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The vnodes-to-tablets migration creates tablet maps that mirror the
vnode layout: one tablet per vnode, preserving token boundaries and
replica placement. However, due to tablet restrictions, the migration
requires vnode tokens to be a power of two and uniformly distributed
across the token ring.
In practice, this restriction is too limiting. Real clusters use
randomly generated tokens and a node's token assignment is immutable.
To solve this problem, prior work (01fb97ee78) has been done to relax
the tablet constraints by allowing arbitrary tablet boundaries, removing
the requirement for power-of-two sizing and uniform distribution.
This patch leverages the relaxed tablet constraints to enable tablet map
creation from arbitrary vnode tokens:
* Removes all token-related constraints.
* Handles wrap-around vnodes. If a vnode wraps (i.e., the highest vnode
token is not `dht::token::last()`), it is split into two tablets:
- (last_vnode_token, dht::token::last()]
- [dht::token::first(), first_vnode_token]
The migration ops guide has been updated to remove the power-of-two
constraint.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
`prepare_for_tablets_migration()` is idempotent; it filters out tables
that already have tablet maps and returns early if no tablet maps need
to be created. However, this precondition is currently misplaced. Move
it higher to skip extra work.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
There are several reasons we want to do that.
One is that it will give us more flexibility in distributing the
load. We can subdivide tablets at any token, and achieve more
evenly-sized tablets. In particular, we can isolate large partitions
into separate tablets.
We can also split and merge incrementally individual tablets.
Currently, we do it for the whole table or nothing, which makes
splits and merges take longer and cause wide swings of the count.
This is not implemented in this PR yet, we still split/merge the whole table.
Another reason is vnode to tablets migration. We now could construct a
tablet map which matches exactly the vnode boundaries, so migration
can happen transparently from CQL-coordinator point of view.
Tablet count is still a power-of-two by default for newly created tables.
It may be different if tablet map is created by non-standard means,
or if per-table tablet option "pow2_count" is set to "false".
build/release/scylla perf-tablets:
Memory footprint for 131k tablets increased from 56 MiB to 58.1 MiB (+3.5%)
Before:
```
Generating tablet metadata
Total tablet count: 131072
Size of tablet_metadata in memory: 57456 KiB
Copied in 0.014346 [ms]
Cleared in 0.002698 [ms]
Saved in 1234.685303 [ms]
Read in 445.577881 [ms]
Read mutations in 299.596313 [ms] 128 mutations
Read required hosts in 247.482742 [ms]
Size of canonical mutations: 33.945053 [MiB]
Disk space used by system.tablets: 1.456761 [MiB]
Tablet metadata reload:
full 407.69ms
partial 2.65ms
```
After:
```
Generating tablet metadata
Total tablet count: 131072
Size of tablet_metadata in memory: 59504 KiB
Copied in 0.032475 [ms]
Cleared in 0.002965 [ms]
Saved in 1093.877441 [ms]
Read in 387.027100 [ms]
Read mutations in 255.752121 [ms] 128 mutations
Read required hosts in 211.202805 [ms]
Size of canonical mutations: 33.954453 [MiB]
Disk space used by system.tablets: 1.450162 [MiB]
Tablet metadata reload:
full 354.50ms
partial 2.19ms
```
Closesscylladb/scylladb#28459
* github.com:scylladb/scylladb:
test: boost: tablets: Add test for merge with arbitrary tablet count
tablets, database: Advertise 'arbitrary' layout in snapshot manifest
tablets: Introduce pow2_count per-table tablet option
tablets: Prepare for non-power-of-two tablet count
tablets: Implement merged tablet_map constructor on top of for_each_sibling_tablets()
tablets: Prepare resize_decision to hold data in decisions
tablets: table: Make storage_group handle arbitrary merge boundaries
tablets: Make stats update post-merge work with arbitrary merge boundaries
locator: tablets: Support arbitrary tablet boundaries
locator: tablets: Introduce tablet_map::get_split_token()
dht: Introduce get_uniform_tokens()
Currently cancellation is logged in get_next_task, but the function is
called by tablets code as well where we do not act upon its result, only
yield to the topology coordinator. But the topology coordinator will not
necessary do the cancellation as well since it can be busy with tablets
migration. As a result cancellation is logged, but not done which is
confusing. Fix it by logging cancellation when it is actually happens.
Fixes https://scylladb.atlassian.net/browse/SCYLLADB-1409Closesscylladb/scylladb#29471
When a replica disconnects during a digest read (e.g., during
decommission), the speculating_read_executor now immediately fires
the pending speculative retry instead of waiting for the timer.
On DISCONNECT, the digest_read_resolver invokes an _on_disconnect
callback set by the executor. The callback cancels the speculate
timer and rearms it to clock_type::now() (lowres_clock::now() =
thread-local memory read, no syscall). The existing timer callback
fires on the next reactor poll with all its logic intact — checking
is_completed(), calling add_wait_targets(1), sending the request,
and incrementing speculative_digest_reads/speculative_data_reads.
The notification is fire-and-forget: on_error() does NOT absorb the
DISCONNECT. The existing error arithmetic in digest_read_resolver
already handles this correctly because _target_count_for_cl accounts
for the speculative target.
For never_speculating_read_executor (no spare target) and
always_speculating_read_executor (all requests sent upfront),
_on_disconnect is never set — no behavior change.
Fixesscylladb/scylladb#26307Closesscylladb/scylladb#29428
all_sibling_tablet_replicas_colocated was using committed ti.replicas to
decide whether sibling tablets are co-located and merge can be finalized.
This caused a false non-co-located window when a co-located pair was moved
by the load balancer: as both tablets migrate together, their del_transition
commits may land in different Raft rounds. After the first commit, ti.replicas
diverge temporarily (one tablet shows the new position, the other the old),
causing all_sibling_tablet_replicas_colocated to return false. This clears
finalize_resize, allowing the load balancer to start new cascading migrations
that delay merge finalization by tens of seconds.
Fix this by using the optimistic replica view (trinfo->next when transitioning,
ti.replicas otherwise) — the same view the load balancer uses for load
accounting — so finalize_resize stays populated throughout an in-flight
migration and no spurious cascades are triggered.
Steps that lead to the problem:
1. Merge is triggered. The load balancer generates co-location migrations
for all sibling pairs that are not yet on the same shard. Some pairs
finish co-location before others.
2. Once all pairs are co-located in committed state,
all_sibling_tablet_replicas_colocated returns true and finalize_resize
is set. Meanwhile the load balancer may have already started a regular
LB migration on one co-located pair (both tablets are stable and the
load balancer is free to move them).
3. The LB migration moves both tablets together (colocated_tablets). Their
two del_transition commits land in separate Raft rounds. After the first
commit, ti.replicas[t1] = new position but ti.replicas[t2] = old position.
4. In this window, all_sibling_tablet_replicas_colocated sees the pair as
NOT co-located, clears finalize_resize, and the load balancer generates
new migrations for other tablets to rebalance the load that the pair
move created.
5. Those new migrations can take tens of seconds to stream, keeping the
coordinator in handle_tablet_migration mode and preventing
maybe_start_tablet_resize_finalization from being called. The merge
finalization is delayed until all those cascaded migrations complete.
Fixes https://scylladb.atlassian.net/browse/SCYLLADB-821.
Fixes https://scylladb.atlassian.net/browse/SCYLLADB-1459.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Closesscylladb/scylladb#29465
Currently the statement returns cluster, partitioner and snitch names by accessing global db::config via database. As the part of an effort to detach components from global db::config, this PR tweaks the statement handler to get the cluster information from some other source. Currently the needed cluster information is stored in different components, but they are all under storage_service umbrella which seems to be a good central source of this truth. Unit test included.
Cleaning components inter-dependencies, not backporting
Closesscylladb/scylladb#29429
* github.com:scylladb/scylladb:
test: Add test_describe_cluster_sanity for DESCRIBE CLUSTER validation
describe_statement: Get cluster info from storage_service
storage_service: Add describe_cluster() method
query_processor: Expose storage_service accessor
RF change of tablet keyspace starts tablet rebuilds. Even if any of the rebuilds is rolled back (because pending replica was excluded), rf change request finishes successfully. In this case we end up with the state of the replicas that isn't compatible with the expected keyspace replication.
Modify topology coordinator so that if it were to be idle, it starts checking if there are any missing replicas. It moves to transition_state::tablet_migration and run required rebuilds.
If a new RF change request encounters invalid state of replicas it fails. The state will be fixed later and the analogical ALTER KEYSPACE statement will be allowed.
Fixes: SCYLLADB-109.
Requires backport to all versions with tablet keyspace rf change.
Closesscylladb/scylladb#28709
* github.com:scylladb/scylladb:
test: add test_failed_tablet_rebuild_is_retried_on_alter
test: add a test to ensure that failed rebuilds are retried
service: fail ALTER KEYSPACE if replicas do not satisfy the replication
service: retry failed tablet rebuilds
service: maybe_start_tablet_migration returns std::optional<group0_guard>
By default it's true, in which case tablet count of the table is
rounded up to a power of two. This option allows lifting this, in
which case the count can be arbitrary. This will allow testing the
logic of arbitrary tablet count.
This is a step towards more flexibility in managing tablets. A
prerequisite before we can split individual tablets, isolating hot
partitions, and evening-out tablet sizes by shifting boundaries.
After this patch, the system can handle tables with arbitrary tablet
count. Tablet allocator is still rounding up desired tablet count to
the nearest power of two when allocating tablets for a new table, so
unless the tablet map is allocated in some other way, the counts will
be still a power of two.
We plan to utilize arbitrary count when migrating from vnodes to
tablets, by creating a tablet map which matches vnode boundaries.
One of the reasons we don't give up on power-of-two by default yet is
that it creates an issue with merges. If tablet count is odd, one of
the tablets doesn't have a sibling and will not be merged. That can
obviously cause imbalance of token space and tablet sizes between
tablets. To limit the impact, this patch dynamically chooses which
tablet to isolate when initiating a merge. The largest tablet is
chosen, as that will minimize imbalance. Otherwise, if we always chose
the last tablet to isolate, its size would remain the same while other
tablets double in size with each odd-count merge, leading to
imbalance. The imbalance will still be there, but the difference in
tablet sizes is limited to 2x.
Example (3 tablets):
[0] owns 1/3 of tokens
[1] owns 1/3 of tokens
[2] owns 1/3 of tokens
After merge:
[0] owns 2/3 of tokens
[1] owns 1/3 of tokens
What we would like instead:
Step 1 (split [1]):
[0] owns 1/3 of tokens
[1] old 1.left, owns 1/6 of tokens
[2] old 1.right, owns 1/6 of tokens
[3] owns 1/3 of tokens
Step 2 (merge):
[0] owns 1/2 of tokens
[1] owns 1/2 of tokens
To do that, we need to be able to split individual tablets, but we're
not there yet.
There are several reasons we want to do that.
One is that it will give us more flexibility in distributing the
load. We can subdivide tablets at any points, and achieve more
evenly-sized tablets. In particular, we can isolate large partitions
into separate tablets.
Another reason is vnode-to-tablet migration. We could construct a
tablet map which matches exactly the vnode boundaries, so migration
can happen transparently from the CQL-coordinator's point of view.
Implementation details:
We store a vector of tokens which represent tablet boundaries in the
tablet_id_map. tablet_id keeps its meaning, it's an index into vector
of tablets. To avoid logarithmic lookup of tablet_id from the token,
we introduce a lookup structure with power-of-two aligned buckets, and
store the tablet_id of the tablet which owns the first token in the
bucket. This way, lookup needs to consider tablet id range which
overlaps with one bucket. If boundaries are more or less aligned,
there are around 1-2 tablets overlapping with a bucket, and the lookup
is still O(1).
Amount of memory used increased, but not significantly relative to old
size (because tablet_info is currently fat):
For 131'072 tablets:
Before:
Size of tablet_metadata in memory: 57456 KiB
After:
Size of tablet_metadata in memory: 59504 KiB
