Today S3 client has well established and well testes (hopefully) http request retry strategy, in the rest of clients it looks like we are trying to achieve the same writing the same code over and over again and of course missing corner cases that already been addressed in the S3 client.
This PR aims to extract the code that could assist other clients to detect the retryability of an error originating from the http client, reuse the built in seastar http client retryability and to minimize the boilerplate of http client exception handling
No backport needed since it is only refactoring of the existing code
Closesscylladb/scylladb#28250
* github.com:scylladb/scylladb:
exceptions: add helper to build a chain of error handlers
http: extract error classification code
aws_error: extract `retryable` from aws_error
- Correct `calc_part_size` function since it could return more than 10k parts
- Add tests
- Add more checks in `calc_part_size` to comply with S3 limits
Fixes: https://scylladb.atlassian.net/browse/SCYLLADB-640
Must be ported back to 2025.3/4 and 2026.1 since we may encounter this bug in production clusters
Closesscylladb/scylladb#28592
* github.com:scylladb/scylladb:
s3_client: add more constrains to the calc_part_size
s3_client: add tests for calc_part_size
s3_client: correct multipart part-size logic to respect 10k limit
Improves performance of deserialization of vector data for calculating similarity functions.
Instead of deserializing vector data into a std::vector<data_value>, we deserialize directly into a std::vector<float>
and then pass it to similarity functions as a std::span<const float>.
This avoids overhead of data_value allocations and conversions.
Example QPS of `SELECT id, similarity_cosine({vector<float, 1536>}, {vector<float, 1536>}) ...`:
client concurrency 1: before: ~135 QPS, after: ~1005 QPS
client concurrency 20: before: ~280 QPS, after: ~2097 QPS
Measured using https://github.com/zilliztech/VectorDBBench (modified to call above query without ANN search)
Fixes https://scylladb.atlassian.net/browse/SCYLLADB-471Closesscylladb/scylladb#28615
Fixes#28398Fixes#28399
When used as path elements in google storage paths, the object names need to be URL encoded. Due to
a.) tests not really using prefixes including non-url valid chars (i.e. / etc)
and
b.) the mock server used for most testing not enforcing this particular aspect,
this was missed.
Modified unit tests to use prefixing for all names, so when running real GS, any errors like this will show.
"Real" GCS also behaves a bit different when listing with pager, compared to mock;
The former will not give a pager token for last page, only penultimate.
Adds handling for this.
Needs backport to the releases that have (though might not really use) the feature, as it is technically possible to use google storage for backup and whatnot there, and it should work as expected.
Closesscylladb/scylladb#28400
* github.com:scylladb/scylladb:
utils/gcp/object_storage: URL-encode object names in URL:s
utils::gcp::object_storage: Fix list object pager end condition detection
It turns out that the cdc driver requires permissions to two additional system tables. This patch adds them to VECTOR_SEARCH_INDEXING and modifies the unit tests. The integration with vector store was tested manually, integration tests will be added in vector-store repository in a follow up PR.
Fixes: SCYLLADB-522
Closesscylladb/scylladb#28519
related PR: https://github.com/scylladb/scylladb/pull/27527
This PR changes test.py logic of parsing boost test cases to use -- --list_json_content
and pass boost labels as pytests markers
using -- --list_json_content is not ideal and currenly require to implement severall [workarounds](https://github.com/scylladb/scylladb/pull/27527#issuecomment-3765499812), but having the ability to support boost labels in pytest is worth it. because now we can apply the tiering mechanism for the boost tests as well
Fixes SCYLLADB-246
Closesscylladb/scylladb#28232
* github.com:scylladb/scylladb:
test: add nightly label
test.py: support boost labels in test.py
add nightly label for test
test_foreign_reader_as_mutation_source
as an example of usinf boost labels pytest as markers
command to test :
./tools/toolchain/dbuild pytest --test-py-init --collect-only -q -m=nightly test/boost
output:
boost/mutation_reader_test.cc::test_foreign_reader_as_mutation_source.debug.1
boost/mutation_reader_test.cc::test_foreign_reader_as_mutation_source.release.1
boost/mutation_reader_test.cc::test_foreign_reader_as_mutation_source.dev.1
Move to `replica/`, drop `flat` from name and drop unused usages as well as unused includes.
Code cleanup, no backport
Closesscylladb/scylladb#28353
* github.com:scylladb/scylladb:
replica/partition_snapshot_reader: remove unused includes
partition_snapshot_reader: remove "flat" from name
mv partition_snapshot_reader.hh -> replica/
This test case was observed to take over 2 minutes to run on CI
machines, contributing to already bloated CI run times.
Disable this test in debug mode. This test checks for memtable flush
being slowed down when compaction can't keep up. So this test needs to
overwhelm the CPU by definition. On the other hand, this is not a
correctness test, there are such tests for the memtable and compaction
already, so it is not critical to run this in debug mode, it is not
expected to catch any use-after-free and such.
Closesscylladb/scylladb#28407
When reads arrive, they have to wait for admission on the reader
concurrency semaphore. If the node is overloaded, the reads will
be queued. They can time out while in the queue, but will not time
out once admitted.
Once the shard is sufficiently loaded, it is possible that most
queued reads will time out, because the average time it takes to
for a queued read to be admitted is around that of the timeout.
If a read times out, any work we already did, or are about to do
on it is wasted effort. Therefore, the patch tries to prevent it
by checking if an admitted read has a chance to complete in time
and abort it if not. It uses the following criteria:
if read's remaining time <= read's timeout when arrived to the semaphore * live updateable preemptive_abort_factor;
the read is rejected and the next one from the wait list is considered.
Fixes https://github.com/scylladb/scylladb/issues/14909
Fixes: SCYLLADB-353
Backport is not needed. Better to first observe its impact.
Closesscylladb/scylladb#21649
* github.com:scylladb/scylladb:
reader_concurrency_semaphore: Check during admission if read may timeout
permit_reader::impl: Replace break with return after evicting inactive permit on timeout
reader_concurrency_semaphore: Add preemptive_abort_factor to constructors
config: Add parameters to control reads' preemptive_abort_factor
permit_reader: Add a new state: preemptive_aborted
reader_concurrency_semaphore: validate waiters counter when dequeueing a waiting permit
reader_concurrency_semaphore: Remove cpu_concurrency's default value
When a shard on a replica is overloaded, it breaks down completely,
throughput collapses, latencies go through the roof and the
node/shard can even become completely unresponsive to new connection
attempts.
When reads arrive, they have to wait for admission on the reader
concurrency semaphore. If the node is overloaded, the reads will
be queued and thus they can time out while being in the queue or during
the execution. In the latter case, the timeout does not always
result in the read being aborted.
Once the shard is sufficiently loaded, it is possible that most
queued reads will time out, because the average time it takes
for a queued read to be admitted is around that of the timeout.
If a read times out, any work we already did, or are about to do
on it is wasted effort. Therefore, the patch tries to prevent it
by checking if an admitted read has a chance to complete in time
and abort it if not. It uses the following cryteria:
if read's remaining time <= read's timeout when arrived to the semaphore * preemptive factor;
the read is rejected and the next one from the wait list is
considered.
The new parameter parametrizes the factor used to reject a read
during admission. Its value shall be between 0.0 and 1.0 where
+ 0.0 means a read will never get rejected during admission
+ 1.0 means a read will immediatelly get rejected during admission
Although passing values outside the interaval is possible, they
will have the exact same effects as they were clamped to [0.0, 1.0].
Fixes#28398
When used as path elements in google storage paths, the object names
need to be URL encoded. Due to a.) tests not really using prefixes including
non-url valid chars (i.e. / etc) and the mock server used for most
testing not enforcing this particular aspect, this was missed.
Modified unit tests to use prefixing for all names, so when run
in real GS, any errors like this will show.
In production environments, we observed cases where the S3 client would repeatedly fail to connect due to DNS entries becoming stale. Because the existing logic only attempted the first resolved address and lacked a way to refresh DNS state, the client could get stuck in a failure loop.
Introduce RR TTL and connection failure retry to
- re-resolve the RR in a timely manner
- forcefully reset and re-resolve addresses
- add a special case when the TTL is 0 and the record must be resolved for every request
Fixes: CUSTOMER-96
Fixes: CUSTOMER-139
Should be backported to 2025.3/4 and 2026.1 since we already encountered it in the production clusters for 2025.3
Closesscylladb/scylladb#27891
* github.com:scylladb/scylladb:
connection_factory: includes cleanup
dns_connection_factory: refine the move constructor
connection_factory: retry on failure
connection_factory: introduce TTL timer
connection_factory: get rid of shared_future in dns_connection_factory
connection_factory: extract connection logic into a member
connection_factory: remove unnecessary `else`
connection_factory: use all resolved DNS addresses
s3_test: remove client double-close
The commit 59faa6d, introduces a new parameter called cpu_concurrency
and sets its default value to 1 which violates the commit fbb83dd that
removes all default values from constructors but one used by the unit
tests.
The patch removes the default value of the cpu_concurrency parameter
and alters tests to use the test dedicated reader_concurrency_semaphore
constructor wherever possible.
consistent_cluster_management is deprecated since scylla-5.2 and no
longer used by Scylladb, so it should not be used by test either.
Closesscylladb/scylladb#28340
These two streams mostly play together. The former provides an input_stream from read from in-memory temporary buffers, the latter wraps it to limit the size of provided temporary buffers. Both are used to test contiguous data consumer, also the buffer_input_stream has a caller in sstables reversing reader.
This PR removes the buffer_input_stream in favor of seastar memory_data_source, and moves the limiting_input_stream into test/lib.
Enanching testing code, not backporting
Closesscylladb/scylladb#28352
* github.com:scylladb/scylladb:
code: Move limiting data source to test/lib
util: Simplify limiting_data_source API
util: Remove buffer_input_stream
test: Use seastar::util::temporary_buffer_data_source in data consumer test
sstables: Use seastar::util::as_input_stream() in mx reader
This compaction group testing is useless because the machinery for it
to work was removed. This was useful in the early tablet days, where
we wanted to test compaction groups directly. Today groups are stressed
and tested on every tablet test.
I see a ~40% reduction time after this patch, since database_test is
one of the most (if not the most) time consuming in boost suite.
Signed-off-by: Raphael S. Carvalho <raphaelsc@scylladb.com>
Closesscylladb/scylladb#28324
The partition snapshot lives in mutation/, however mutation/ is a lower
level concept than a mutation reader. The next best place for this
reader is the replica/ directory, where the memtable, its main user,
also lives.
Also move the code to the replica namespace.
test/boost/mvcc_test.cc includes this header but doesn't use anything
from it. Instead of updating the include path, just drop the unused
include.
Only two tests use it now -- the limit-data-source-test iself and a test
that validates continuous_data_consumer template.
Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
The source maintains "limit generator" -- a function that returns the
maximum size of bytes to return from the next buffer.
Currently all callers just return constant numbers from it. Passing a
function that returns non-constant one can, probably, be used for a
fuzzy test, but even the limiting-data-source-test itself doesn't do it,
so what's the point...
Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
The test creates buffer_data_source_impl and wraps it with limiting data
source. The former data_source duplicates the functionality of the
existing seastar temporary_buffer_data_source.
This patch makes the test code use seastar facility. The
buffer_data_source_impl will be removed soon.
Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
Previously we only inspected std::system_error inside
std::nested_exception to support a specific TLS-related failure
mode. However, nested exceptions may contain any type, including
other restartable (retryable) errors. This change unwraps one
nested exception per iteration and re-applies all known handlers
until a match is found or the chain is exhausted.
Closesscylladb/scylladb#28240
Allows other topology operations to execute while tablets are being
drained on decommission. In particular, bootstrap on scale-out. This
is important for elasticity.
Allows multiple decommission/removenode to happen in parallel, which
is important for efficiency.
Flow of decommission/removenode request:
1) pending and paused, has tablet replicas on target node.
Tablet scheduler will start draining tablets.
2) No tablets on target node, request is pending but not paused
3) Request is scheduled, node is in transition
4) Request is done
Nodes are considered draining as soon as there is a leave or remove
request on them. If there are tablet replicas present on the target
node, the request is in a paused state and will not be picked by
topology coordinator. The paused state is computed from topology state
automatically on reload.
When request is not paused, its execution starts in
write_both_read_old state. The old tablet_draining state is not
entered (it's deprecated now).
Tablet load balancing will yield the state machine as soon as some
request is no longer paused and ready to be scheduled, based on
standard preemption mechanics.
Fixes#21452Closesscylladb/scylladb#24129
* https://github.com/scylladb/scylladb:
docs: Document parallel decommission and removenode and relevant task API
test: Add tests for parallel decommission/removenode
test: util: Introduce ensure_group0_leader_on()
test: tablets: Check that there are no migrations scheduled on draining nodes
test: lib: topology_builder: Introduce add_draining_request()
topology_coordinator, tablets: Fail draining operations when tablet migration fails due to critical disk utilization
tablets: topology_coordinator: Refactor to propagate reason for migration rollback
tablet_allocator: Skip co-location on draining nodes
node_ops: task_manager_module: Populate entity field also for active requests
tasks: node_ops: Put node id in the entity field
tasks, node_ops: Unify setting of task_stats in get_status() and get_stats()
topology: Protect against empty cancelation reason
tasks, topology: Make pending node operations abortable
doc: topology-over-raft.md: Fix diagram for replacing, tablet_draining is not engaged
raft_topology, tablets: Drain tablets in parallel with other topology operations
virtual_tables: Show draining and excluded fields in system.cluster_status and system.load_by_node
locator: topology: Add "draining" flag to a node
topology_coordinator: Extract generate_cancel_request_update()
storage_service: Drop dependency in topology_state_machine.hh in the header
locator: Extract common code in assert_rf_rack_valid_keyspace()
topology_coordinator, storage_service: Validate node removal/decommission at request submission time
Adds a "sstables" array member to manifest.json.
For each sstables, keep the following metadata:
id - a uuid for the sstable (the sstable identifier
if the use-sstable-identifier option was used, otherwise
the sstable uuid generation)
toc_name - the name of the TOC.txt file
data_size and index_size - in bytes
first_token and last_token - of the sstable first and last keys.
Fixes: SCYLLADB-196
Signed-off-by: Benny Halevy <bhalevy@scylladb.com>
Add a table member to manifest.json with the keyspace_name,
table_name, table_id, tablets_type, and, for tablets-enabled tables, get
tablet_count on each shard and write the minimum to manifest.json.
For vnodes-based tables, tablet_count=0.
For now, `tablets_type` may be either `none` for vnodes tables, or
`powof2` for tablets tables. In the future, when we support arbitrary
tablt boundaries, this will be reflected here, and it is likely we
would backup the whole tablets map sperately to get all tablet boundaries.
Fixes SCYLLADB-195
Signed-off-by: Benny Halevy <bhalevy@scylladb.com>
Add metadata about the node: host_id, datacenter, and rack.
This enables dc- or rack- aware restore.
Today this information is "encoded" into the snapshot hierarchy
prefixes, but if all manifest files would be stored in a flat
directory, we'd need to encode that metadata in the object name,
but it'd be better for the manifest contents to be self descriptive.
Signed-off-by: Benny Halevy <bhalevy@scylladb.com>
Add metadata about the manifest itself:
A version and the manifest scope (currently "node",
but in the future, may also be "shard", or "tablet")
Signed-off-by: Benny Halevy <bhalevy@scylladb.com>
Validate the manifest.json format by loading it using rjson::parse
and then validate its contents to ensure it lists exactly the
SSTables present in the snapshot directory.
Signed-off-by: Benny Halevy <bhalevy@scylladb.com>
In PR 5b6570be52 we introduced the config option `sstable_compression_user_table_options` to allow adjusting the default compression settings for user tables. However, the new option was hooked into the CQL layer and applied only to CQL base tables, not to the whole spectrum of user tables: CQL auxiliary tables (materialized views, secondary indexes, CDC log tables), Alternator base tables, Alternator auxiliary tables (GSIs, LSIs, Streams).
This gap also led to inconsistent default compression algorithms after we changed the option’s default algorithm from LZ4 to LZ4WithDicts (adf9c426c2).
This series introduces a general “schema initializer” mechanism in `schema_builder` and uses it to apply the default compression settings uniformly across all user tables. This ensures that all base and aux tables take their default compression settings from config.
Fixes#26914.
Backport justification: LZ4WithDicts is the new default since 2025.4, but the config option exists since 2025.2. Based on severity, I suggest we backport only to 2025.4 to maintain consistency of the defaults.
Closesscylladb/scylladb#27204
* github.com:scylladb/scylladb:
db/config: Update sstable_compression_user_table_options description
schema: Add initializer for compression defaults
schema: Generalize static configurators into schema initializers
schema: Initialize static properties eagerly
db: config: Add accessor for sstable_compression_user_table_options
test: Check that CQL and Alternator tables respect compression config
This patch changes the layout of user-facing scheduling groups from
/
`- statement
`- sl:default
`- sl:*
`- other groups (compaction, streaming, etc.)
into
/
`- user (supergroup)
`- statement
`- sl:default
`- sl:*
`- other groups (compaction, streaming, etc.)
The new supergroup has 1000 static shares and is name-less, in a sense
that it only have a variable in the code to refer to and is not exported
via metrics (should be fixed in seastar if we want to).
The moved groups don't change their names or shares, only move inside
the scheduling hierarchy.
The goal of the change is to improve resource consumption of sl:*
groups. Right now activities in low-shares service levels are scheduled
on-par with e.g. streaming activity, which is considered to be low-prio
one. By moving all sl:* groups into their own supergroup with 1000
shares changes the meaning of sl:* shares. From now on these shares
values describe preirities of service level between each-other, and the
user activities compete with the rest of the system with 1000 shares,
regardless of how many service levels are there.
Unit tests keep their user groups under root supergroup (for simplicity)
Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
Closesscylladb/scylladb#28235
The API contract in partition_version.hh states that when dealing with
evictable entries, a real cache tracker pointer has to be passed to all
methods that ask for it. The nonpopulating reader violates this, passing
a nullptr to the snapshot. This was observed to cause a crash when a
concurrent cache read accessed the snapshot with the null tracker.
A reproducer is included which fails before and passes after the fix.
Fixes: #26847Closesscylladb/scylladb#28163
Cassandra changed their system tables in 3.0. We migrated to the new system table layout in 2017, in ScyllaDB 2.0.
System tables introduced in Cassandra 3.0, as well as the 3.0 variant of pre-existing system tables were added to the db::system_table::v3 namespace.
We ended up adding some new ScyllaDB-only system tables to this namespace as well.
As the dust settled, most of the v3 system tables ended up being either simple aliases to non-v3 tables, or new tables.
Either way, the codebase uses just one variant of each table for a long time now the v3:: distinction is pointless.
Remove the v3 namespace and unify the table listing under the top-level db::system_keyspace scope.
Code cleanup, no backport
Closesscylladb/scylladb#28146
* github.com:scylladb/scylladb:
db/system_keyspace: move remining tables out of v3 keyspace
db/system_keyspace: relocate truncated() and commitlog_cleanups()
db/system_keyspace: drop v3::local()
db/system_keyspace: remove duplicate table names from v3
The last remining tables in the v3 keyspace are those that are genuinely
distinct -- added by Cassandra 3.0 or >= ScyllaDB 2.0.
Move these out of the v3 keyspace too, with this the v3 keyspace is
defunct and removed.
The loop that unwraps nested exception, rethrows nested exception and saves pointer to the temporary std::exception& inner on stack, then continues. This pointer is, thus, pointing to a released temporary
Closesscylladb/scylladb#28143
reader_permit::release_base_resources() is a soft evict for the permit:
it releases the resources aquired during admission. This is used in
cases where a single process owns multiple permits, creating a risk for
deadlock, like it is the case for repair. In this case,
release_base_resources() acts as a manual eviction mechanism to prevent
permits blockings each other from admission.
Recently we found a bad interaction between release_base_resources() and
permit eviction. Repair uses both mechanism: it marks its permits as
inactive and later it also uses release_base_resources(). This partice
might be worth reconsidering, but the fact remains that there is a bug
in the reader permit which causes the base resources to be released
twice when release_base_resources() is called on an already evicted
permit. This is incorrect and is fixed in this patch.
Improve release_base_resources():
* make _base_resources const
* move signal call into the if (_base_resources_consumed()) { }
* use reader_permit::impl::signal() instead of
reader_concurrency_semaphore::signal()
* all places where base resources are released now call
release_base_resources()
A reproducer unit test is added, which fails before and passes after the
fix.
Fixes: #28083Closesscylladb/scylladb#28155
In case of decommission, it's not desirable because it's less urgent.
In case of removenode, it leads to failure of removenode operation
because scheduled co-locating migration will fail if the destination
is on the excluded node, and this failure will be interpreted as drain
failure and coordinator will cancel the request.
Not a problem before "parallel decommission" because this failure is
only a streaming failure, not a barrier failure, so exception doesn't
escape into the catch clause in transition stage handler, and the
migration is simply rolled back. Once draining happens in the tablet
migration track, streaming failure will be interpreted as drain
failure and cancel the request.
The `make_key` lambda erroneously allocates a fixed 8-byte buffer
(`sizeof(s.size())`) for variable-length strings, potentially causing
uninitialized bytes to be included. If such bytes exist and they are
not valid UTF-8 characters, deserialization fails:
```
ERROR 2026-01-16 08:18:26,062 [shard 0:main] testlog - snapshot_list_contains_dropped_tables: cql env callback failed, error: exceptions::invalid_request_exception (Exception while binding column p1: marshaling error: Validation failed - non-UTF8 character in a UTF8 string, at byte offset 7)
```
Fixes#28195.
Signed-off-by: Nikos Dragazis <nikolaos.dragazis@scylladb.com>
Closesscylladb/scylladb#28197
Continuing the read once it is aborted (e.g. due to timeout) is a waste
of resources, as the produced results will be discarded.
Poll the permit's abort exception in the memtable and cache reader's
fill_buffer(). This results in one poll per buffer filled (8KB of data).
We already have similar poll for sstable readers, as disk reads are
usually much heavier and therefore it is more important to stop them
ASAP after abort. Cache and memtable reads are usually quick but not
always, hence it is important to also have polling in the cache and
memtable readers.
Refs: #11469Fixes: #28148Closesscylladb/scylladb#28149
