a11::stores::SQLiteChunkStoreOptions Struct Reference

Storage layout policy for SQLiteChunkStore. More...

#include <cpp/a11/stores/sqlite_chunk_store.h>

Public Member Functions
absl::Status	Validate () const
	Validate the threshold, owner id and durations.

Static Public Member Functions
static absl::StatusOr< SQLiteChunkStoreOptions >	FromEnvironment ()
	Read storage policy from the A11_SQLITE_CHUNK_STORE_* environment values.

Public Attributes
size_t	inline_data_threshold = 128 * 1024
	Payloads strictly larger than this move out of the row into a blob file.
std::string	owner_id
	Recorded on the node row as its owner.
internal::SqliteSynchronous	synchronous
	Durability level applied with PRAGMA synchronous.
absl::Duration	cross_process_poll_interval = absl::ZeroDuration()
	How often to look for commits made by other processes, or zero to disable.
absl::Duration	blob_grace_period = absl::Hours(1)
	How long an unreferenced blob must survive before a sweep may remove it.

Friends
bool	operator== (const SQLiteChunkStoreOptions &, const SQLiteChunkStoreOptions &)=default

Detailed Description

Storage layout policy for SQLiteChunkStore.

Member Function Documentation

FromEnvironment()

absl::StatusOr< SQLiteChunkStoreOptions > a11::stores::SQLiteChunkStoreOptions::FromEnvironment ( )

static

Read storage policy from the A11_SQLITE_CHUNK_STORE_* environment values.

Validate()

absl::Status a11::stores::SQLiteChunkStoreOptions::Validate

(

)

const

Validate the threshold, owner id and durations.

Friends And Related Symbol Documentation

operator==

bool operator== ( const SQLiteChunkStoreOptions &  , const SQLiteChunkStoreOptions &    )

friend

Member Data Documentation

blob_grace_period

absl::Duration a11::stores::SQLiteChunkStoreOptions::blob_grace_period = absl::Hours(1)

How long an unreferenced blob must survive before a sweep may remove it.

cross_process_poll_interval

absl::Duration a11::stores::SQLiteChunkStoreOptions::cross_process_poll_interval = absl::ZeroDuration()

How often to look for commits made by other processes, or zero to disable.

SQLite's change hooks are per-connection, so a writer in another process cannot wake a reader parked in this one. Zero – the default – means in-process notification only, which is exact and never polls.

inline_data_threshold

size_t a11::stores::SQLiteChunkStoreOptions::inline_data_threshold = 128 * 1024

Payloads strictly larger than this move out of the row into a blob file.

Keeping small payloads inline avoids a file per chunk; keeping large ones out of the row keeps the database page cache useful and the WAL small.

owner_id

std::string a11::stores::SQLiteChunkStoreOptions::owner_id

Recorded on the node row as its owner.

Ownership carries no enforcement yet; it exists so nodes can be attributed and filtered.

synchronous

internal::SqliteSynchronous a11::stores::SQLiteChunkStoreOptions::synchronous

Initial value:

=
      internal::SqliteSynchronous::kNormal

Durability level applied with PRAGMA synchronous.

---

The documentation for this struct was generated from the following files:

• cpp/a11/stores/sqlite_chunk_store.h
• cpp/a11/stores/sqlite_chunk_store.cc