Replay the WAL log, if any, at startup.
Op replay is performed synchronously, during initialisation of the
ingester2 instance, and passes all ops through the "normal" write path
that the system uses once replay is complete, minus the WAL writer layer
- this helps to DRY the write path and minimise different behaviours.
Now partition cache entries are smaller, the number of entries held in
memory can be increased - this now uses ~2MiB of memory and drains the
cache during execution, amortising to 0.
This commit removes the invariant asserts of monotonicity carried over
from the "ingester" crate - ingester2 does not define any ordering of
writes within the system.
This commit also removes the SequenceNumberRange as it is no longer
useful to indirectly check the equality of two sets of ops -
non-monotonic writes means overlapping ranges does not guarantee a full
set of overlapping operations (gaps may be present). Likewise bounding
values (such as "min persisted") provide no useful meaning in an
out-of-order system.
Document the arbitrary reordering of concurrent writes in an ingester,
and the potential divergence of WAL entries / buffered state.
Also documents that in ingester2, sequence numbers only identify writes,
not their ordering.
Adds WalSink, an implementation of the DmlSink trait that commits DML
operations to the write-ahead log before passing the DML op into the
decorated inner DmlSink chain.
By structuring the WAL logic as a decorator, the chain of inner DmlSink
handlers within it are only ever invoked after the WAL commit has
successfully completed, which keeps all the WAL commit code in a
singly-responsible component. This also lets us layer on the WAL commit
logic to the DML sink chain after replaying any existing WAL files,
avoiding a circular WAL mess.
The application-logic level WAL code abstracts over the underlying WAL
implementation & codec through the WalAppender trait. This decouples the
business logic from the WAL implementation both for testing purposes,
and for trying different WAL implementations in the future.
* chore: Update Datafusion and arrow/arrow-flight/parquet to `28.0.0`
* chore: Update thrift to 0.17
* fix: use workspace arrow-flight in ingester2
* chore: Update for API changes
* fix: test
* chore: Update hakari
* chore: Update hakari again
* chore: Update trace_exporters to latest thrift
* fix: update test
Co-authored-by: kodiakhq[bot] <49736102+kodiakhq[bot]@users.noreply.github.com>
Sequence all gRPC write requests to (internally) order the resulting DML
operations.
These sequence numbers are assigned from a timestamp oracle and passed
through to the downstream DmlSink implementers.
This commit implements the QueryExec trait for the BufferTree, allow it
to be queried for the partition data it contains. With this change, the
BufferTree now provides "read your writes" functionality.
Notably the implementation streams the contents of individual partitions
to the caller on demand (pull-based execution), deferring acquiring the
partition lock until actually necessary and minimising the duration of
time a strong reference to a specific RecordBatch is held in order to
minimise the memory overhead.
During query execution a client sees a consistent snapshot of
partitions: once a client begins streaming the query response, incoming
writes that create new partitions do not become visible. However
incoming writes to an existing partition that forms part of the snapshot
set become visible iff they are ordered before the acquisition of the
partition lock when streaming that partition data to the client.
Allow the return type of the QueryExec trait's query_exec() method to be
parametrised by the implementer.
This allows the trait to be reused across different data sources that
return differing concrete types.
Changes the query code (taken from the ingester crate) to stream data
for query execution, tidy up unnecessary Result types and removing
unnecessary indirection/boxing.
Previously the query data sourcing would collect the set of RecordBatch
for a query response during execution, prior to sending the data to the
caller. Any data that was dropped or modified during this time meant the
underlying ref-counted data could not be released from memory until all
outstanding queries referencing it completed. When faced with multiple
concurrent queries and ongoing ingest, this meant multiple copies of
data could be held in memory at any one time.
After this commit, data is streamed to the user, minimising the duration
of time a reference to specific partition data is held, and therefore
eliminating the memory overhead of holding onto all the data necessary
for a query for as long as the client takes to read the data.
When combined with an upcoming PR to stream RecordBatch out of the
BufferTree, this should provide performant query execution with minimal
memory overhead, even for a maliciously slow reading client.
Implement a QueryExec decorator that emits named tracing spans covering
the inner delegate's query_exec() execution.
Captures the result, emitting the error string in the span on failure.
This commit implements the gRPC direct-write RPC interface (largely
copied from the ingester crate), and adds a much improved RPC query
handler.
Compared to the ingester crate, the query API is now split into two
defined halves - the API handler side, and types necessary to support it
(server/grpc/query.rs) and the Ingester query execution side (a stub in
query/exec.rs). These two halves maintain a separation of concerns, and
are interfaced by an abstract QueryExec trait (in query/trait.rs).
I also added the catalog RPC interface as it is currently exposed on the
ingester, though I am unsure if it is used by anything.
This commit also introduces the "init" module, and the
IngesterRpcInterface trait within it. This trait forms the public
ingester2 crate API, defining the complete set of methods external
crates can expect to utilise in a stable, unchanging and decoupled way.
The IngesterRpcInterface trait also serves as a method of type-erasure
on the underlying handler implementations, avoiding the need to
expose/pub the types, abstractions, and internal implementation details
of the ingester to external crates.
Adds an ingester2 crate to hold the MVP of the Kafkaless project.
This was necessary due to the tight coupling of the ingester internals
with tests in external crates, and eases the parallel development of two
version of the ingester.
This commit contains various changes from the "ingester" crate, mostly
removing the concept/references to a "shard" or "ShardId" where
possible.
This commit does not copy over all of the "ingester" crate - only those
components that are definitely needed. I will drag across more as
functionality is implemented.
Dom Dwyer
kodiakhq[bot]
Andrew Lamb
dependabot[bot]
CircleCI[bot]