1. Only call base_ck = generate_base_key_from_index_pk<...
if the base schema has a clustering key.
2. Only call command->slice.set_range(*_schema, base_pk, ...
if the base schema has a clustering key,
otherwise just create an open ended range.
Proposed-by: Piotr Sarna <sarna@scylladb.com>
Signed-off-by: Benny Halevy <bhalevy@scylladb.com>
When called from indexed_table_select_statement::do_execute_base_query,
old_paging_state->get_clustering_key() may return un-engaged
optional<clustering_key>. Dereferencing it unconditionally crashes
scylla as seen in https://github.com/scylladb/scylla/issues/5794Fixes#5794
Signed-off-by: Benny Halevy <bhalevy@scylladb.com>
De-pointerize cql3 code APIs further: change some call sites
to pass `schema` as const-ref instead of `shared_ptr`.
Affected functions known to be expecting always non-null
pointer to schema and don't store or pass the pointer somewhere
else, assuming it's safe to give them just a reference.
Tests: unit(dev, debug)
Signed-off-by: Pavel Solodovnikov <pa.solodovnikov@scylladb.com>
Message-Id: <20200218142338.69824-1-pa.solodovnikov@scylladb.com>
This patch fixes a bug that appears because of an incorrect interaction
between counters and hinted handoff.
When a counter is updated on the leader, it sends mutations to other
replicas that contain all counter shards from the leader. If
consistency level is achieved but some replicas are unavailable, a hint
with mutation containing counter shards is stored.
When a hint's destination node is no longer its replica, it is
attempted to be sent to all its current replicas. Previously, if the
cluster did not have the feature HINTED_HANDOFF_SEPARATE_CONNECTION
enabled, storage_proxy::mutate function would be used for the purpose of
sending the hint. It was incorrect because that function treats
mutations for counter tables as mutations containing only a delta (by
how much to increase/decrease the counter). These two types of mutations
have different serialization format, so in this case a "shards" mutation
is reinterpreted as "delta" mutation, which can cause data corruption to
occur.
This patch fixes the case when HINTED_HANDOFF_SEPARATE_CONNECTION is
disabled, and uses storage_proxy::mutate_internal, which treats "shards"
mutation as regular mutations - which is the correct behavior.
Refs #5833.
Tests: unit(dev)
Refs #817
Truncation is potentially long. It has its own timeout in storage
proxy/rpc. This value should probably also be higher than default
timeout.
Message-Id: <20200218135926.26522-1-calle@scylladb.com>
This patch adds additional tests for the ReturnValues feature to make the
test even more comprehensive. As this feature is not yet implemented in
Alternator (see issue #5053), all tests XFAIL on Alternator - except two
tests for the trivial "NONE" mode which is already supported. As usual
all tests pass on DynamoDB.
This patch also splits the tests for the ReturnValues parameter in the
UpdateItem operation into multiple tests, each testing one of the different
modes which DynamoDB supports - NONE, ALL_OLD, UPDATED_OLD, ALL_NEW and
UPDATED_NEW. The separate tests will be useful if we implement this feature
incrementally - so the separate modes can be tested separately.
Refs #5053.
Signed-off-by: Nadav Har'El <nyh@scylladb.com>
Message-Id: <20200218085618.5584-1-nyh@scylladb.com>
This patch adds a new document, docs/protocols.md, about all the different
protocols which Scylla supports - and the different ports which they use.
This includes Scylla's internal protocol, user-facing protocols (CQL, Thrift,
DynamoDB, Redis, JMX) and things inbetween (REST API, Prometheus).
I wrote this document after being frustrated that when I see a port number
(e.g., "7000") or a port option name (e.g., "storage_port") it's hard to
figure out what they actually are - or why they are given such strange
names. The intention is that this file can easily be searched for option
names, for more familiar names (e.g., "CQL"), and a reader can get the
whole story - including some pointers to relevant part of the code (this
part of the document can be improved further - in this version this only
exists for the internal protocol).
Signed-off-by: Nadav Har'El <nyh@scylladb.com>
Message-Id: <20200217172049.25510-1-nyh@scylladb.com>
* seastar c7c249f67d...2b51022073 (8):
> dns_test: Test with seastar.io instead of www.google.com
> sharded: fix move constructor for peering_sharded_service
Fixes#5814.
> tests: Delete Seastar.dist
> reactor: distinguish structs from classes when befriending
> util/tuple_utils.hh: avoid redundant move
> io_request: do not include fmt/format.h
> reactor: cleanup write_some leftover
> posix: change the signature of accept/try_accept
"
This change set is comprised of several unrelated patches regarding
some cleanups in cql3 layer code.
Most of the changes are aimed at eliminating superfluous `shared_ptr`
usages. In places where it can be safely assumed that objects passed
to the function are considered non-null and constant, these places
were adjusted to use passing as const ref instead.
Other changes incude eliminating unused arguments at some functions
and replacing usages of `shared_ptr<service::pager::paging_state>`
to use `lw_shared_ptr` instead, since `pager::paging_state` is final.
Tests: unit(dev, debug)
"
* 'feature/cql_cleanups_4' of https://github.com/ManManson/scylla:
cql3: minor sweeps through the cql layer code to reduce shared_ptrs count
cql3: change some function signatures to accept const references
cql3: change signatures of several functions to return crefs instead of pointers
cql3: remove unused argument at functions::castas_functions::get
paging_state: switch from shared_ptr to lw_shared_ptr
When replaying a hint with a destination node that is no longer in the
cluster, it will be sent with cl=ALL to all its new replicas. Before
this patch, the MUTATION verb was used, which causes such hints to be
handled on the same connection and with the same priority as regular
writes. This can cause problems when a large number of hints is
orphaned and they are scheduled to be sent at once. Such situation
may happen when replacing a dead node - all nodes that accumulated hints
for the dead node will now send them with cl=ALL to their new replicas.
This patch changes the verb used to send such hints to HINT_MUTATION.
This verb is handled on a separate connection and with streaming
scheduling group, which gives them similar priority to non-orphaned
hints.
Refs: #4712
Tests: unit(dev)
" from Botond
Nodetool scrub rewrites all sstables, validating their data. If corrupt
data is found the scrub is aborted. If the skip-corrupted flag is set,
corrupt data is instead logged (just the keys) and skipped.
The scrubbing algorithm itself is fairly simple, especially that we
already have a mutation stream validator that we can use to validate the
data. However currently scrub is piggy-backed on top of cleanup
compaction. To implement this flag, we have to make scrub a separate
compaction type and propagate down the flag. This required some
massaging of the code:
* Add support for more than two (cleanup or not) compaction types.
* Allow passing custom options for each compaction type.
* Allow stopping a compaction without the manager retrying it later.
Additionally the validator itself needed some changes to allow different
ways to handle errors, as needed by the scrub.
Fixes: #5487
* https://github.com/denesb/nodetool-scrub-skip-corrupted/v7:
table: cleanup_sstables(): only short-circuit on actual cleanup
compaction: compaction_type: add Upgrade
compaction: introduce compaction_options
compaction: compaction_descriptor: use compaction options instead of
cleanup flag
compaction_manager: collect all cleanup related logic in
perform_cleanup()
sstables: compaction_stop_exception: add retry flag
mutation_fragment_stream_validator: split into low-level and
high-level API
compaction: introduce scrub_compaction
compaction_manager: scrub: don't piggy-back on upgrade_sstables()
test: sstable_datafile_test: add scrub unit test
CDC can get all it needs from a config and does not need
partitioner.
For base table specific operations CDC is using partitioner
from that table (obtained with schema::get_partitioner).
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
All the places that use partitioner have been switched
to not use global partitioner any more and we can stop
setting it in this test.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
random_schema already has a _schema field which in turn
has a get_partitioner() function. Store partitioner
in random_schema is redundant.
At the moment all uses of random_schema are based on
default partitioner so it is not necessary to set it
explicitly. If in the future we need random_schema to
work with other partitioners we will add the constructor
back and fix the creation of _schema to contain it. It's
not needed now though.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
parse functions now take const schema& which allows
them to reach a partitioner. It's safe to take schema
by const& because the only caller takes the schema
from an sstable object.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
and replace all calls to dht::global_partitioner().get_token
dht::get_token is better because it takes schema and uses it
to obtain partitioner instead of using a global partitioner.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
and replace all dht::global_partitioner().decorate_key
with dht::decorate_key
It is an improvement because dht::decorate_key takes schema
and uses it to obtain partitioner instead of using global
partitioner as it was before.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
Take const schema& as a parameter of shard_of and
use it to obtain partitioner instead of calling
global_partitioner().
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
ring_position_exponential_sharder calls global_partitioner
in one constructor. Luckily the constructor is never used so
we can remove that constructor.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
This requires a change in a repair that uses
selective_token_range_sharder.
Repair performs operation on a set of tables. We will have to
make sure that all of that tables use the same partitioner.
This is achieved by adding a check to a repair_info constructor.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
Remove ring_position_range_sharder(nonwrapping_range<ring_position>)
which calls another constructor with partitioner obtained with
dht::global_partitioner().
Fix all the places the removed constructor was used and obtain
partitioner from schema instead.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
i_partitioner.hh is widely included while sharders are used
only in 6 places so there's no need to include them in
the whole codebase.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
The next patch is moving sharders to a separate header.
ring_position_exponential_vector_sharder is not used anywhere
so instead of just silently removing it with the move, this
commit is separated to make it clear the class is removed.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
The plan is to remove dht::global_partitioner()
and use schema::get_partitioner() instead.
This will allow a usage of per schema/table partitioner
instead of a single global partitioner everywhere.
Initially schema::get_partitioner will call
dht::global_partitioner. After all the calls
to dht::global_partitioner are switched to
schema::get_partitioner, the ability to set per schema
partitioner will be implemented.
Signed-off-by: Piotr Jastrzebski <piotr@scylladb.com>
