cubdb
cubdb v2.0.2

View Source Changelog

For versions before v1.0.0, the changes are reported in the GitHub releases.

Since v1.0.0, CubDB follows semantic versioning, and reports changes here.

v2-0-2-2023-01-01

 v2.0.2 (2023-01-01)

Bug fixes:

  • Better exception in case of file errors (such as :efbig)

v2-0-1-2022-08-30

 v2.0.1 (2022-08-30)

Various bug fixes:

  • Avoid unnecessary commits in the compaction catching up phase
  • More informative error when running out of disk space
  • Avoid leaking processes when stopping CubDB during a compaction

v2-0-0-2022-06-28

 v2.0.0 (2022-06-28)

Version 2.0.0 brings better concurrency, atomic transactions with arbitrary operations, zero cost read-only snapshots, database backup, and more, all with a simpler and more scalable internal architecture.

Refer to the upgrade guide for how to upgrade from previous versions.

  • [breaking] The functions CubDB.get_and_update/3, CubDB.get_and_update_multi/3, and CubDB.select/2 now return directly result, instead of a {:ok, result} tuple.
  • [breaking] CubDB.get_and_update_multi/4 does not take an option argument anymore, making it CubDB.get_and_update_multi/3. The only available option used to be :timeout, which is not supported anymore.
  • [breaking] Remove the :timeout option on CubDB.select/2. This is part of a refactoring and improvement that moves read operations from an internally spawned Task to the client process. This makes the :timeout option unnecessary: by stopping the process calling CubDB, any running read operation by that process is stopped.
  • [breaking] CubDB.select/2 now returns a lazy stream that can be used with functions in Enum and Stream. This makes the :pipe and :reduce options unnecessary, so those options were removed.
  • Add CubDB.snapshot/2, CubDB.with_snapshot/2 and CubDB.release_snapshot/1 to get zero cost read-only snapshots of the database. The functions in CubDB.Snapshot allow to read from a snapshot.
  • Add CubDB.transaction/2 to perform multiple write (and read) operations in a single atomic transaction. The functions in CubDB.Tx allow to read and write inside a transaction.
  • Add CubDB.back_up/2 to produce a database backup. The backup process does not block readers or writers, and is isolated from concurrent writes.
  • Add CubDB.halt_compaction/1 to stop any running compaction operation
  • Add CubDB.compacting?/1 to check if a compaction is currently running
  • Move read and write operations to the caller process as opposed to the CubDB server process.
  • Improve concurrency of read operations while writing

v1-1-0-2021-10-14

 v1.1.0 (2021-10-14)

  • Add clear/1 function to atomically delete all entries in the database

v1-0-0-2021-06-24

 v1.0.0 (2021-06-24)

breaking-changes-from-v0-17-0

 Breaking changes from v0.17.0:

  • Better defaults:
    • auto_file_sync now defaults to true (slower but durable)
    • auto_compact now defaults to true
  • Functions select/2 and get_and_update_multi/4 now take the timeout as an option instead of an additional argument

other-changes-from-v0-17-0

 Other changes from v0.17.0:

  • Better internal handling of timeouts that ensures cleanup of resources on the callee side
  • Added put_and_delete_multi/3 to atomically put and delete entries
  • Added put_new/3 to put an entry only if the key does not exist yet
  • More efficient implementation of put_multi/2 and delete_multi/2
  • Function get_multi/2 does not block writers
  • Fix race condition during compaction
  • Function get_and_update/3 avoids unnecessary disk writes (including the transaction header) when the value is unchanged
  • Remove caller timeout on put_and_delete_multi/3 and put_multi/2, consistently with the other functions.
  • Remove default GenServer timeouts
  • Fix process (and file descriptor) leak upon compaction
  • Fix cubdb_file?/1 regexp, making it stricter