Skip to content
Snippets Groups Projects
Select Git revision
  • morozov/varbinary
  • sd/wf_fixups protected
  • funbringer/fix-tuple-validate-raw
  • sd/c813750b7
  • funbringer/2.11.5-extend-op-result-row
  • 2.11.5-picodata default protected
  • gmoshkin/fix-window-in-subquery protected
  • sd/fix_r10
  • sd/window
  • sd/fix_sq
  • funbringer/remove-iproto-raft-rollback
  • sd/fix_window_sq_panic
  • sd/64-bit
  • picodata-25.1
  • funbringer/2.11.5-fix-select-param-having
  • ekhamitov/sd/window-v2
  • vshunkov/fix_path protected
  • vshunkov/ci_sock
  • dec_vardir protected
  • vshunkov/fix_checkpatch
  • 2.11.5 protected
  • 2.11.2.159 protected
  • 2.11.4 protected
  • 2.11.2.155 protected
  • 2.11.2.148 protected
  • 2.11.2.145 protected
  • 2.11.2.137 protected
  • 2.11.2.136 protected
  • 2.11.2.135 protected
  • 2.11.3 protected
  • 2.11.2.109 protected
  • 2.11.2.108 protected
  • 2.11.2.105 protected
  • 2.11.2.98 protected
  • 2.11.2.97 protected
  • 2.11.2.96 protected
  • 2.11.2.92 protected
  • 2.11.2 protected
  • 3.0.0-beta1 protected
  • 2.11.0.91 protected
40 results
Compare
user avatar
box: update synchro quorum in on_commit trigger instead of on_replace
Georgiy Lebedev authored 
Currently, we update the synchronous replication quorum from the
`on_replace` trigger of the `_cluster` space when registering a new
replica. However, during the join process, the replica cannot ack its own
insertion into the `_cluster` space. In the scope of #9723, we are going to
enable synchronous replication for most of the system spaces, including the
`_cluster` space. There are several problems with this:

1. Joining a replica to a 1-member cluster without manual changing of
quorum won't work: it is impossible to commit the insertion into the
`_cluster` space with only 1 node, since the quorum will equal to 2 right
after the insertion.

2. Joining a replica to a 3-member cluster may fail: the quorum will become
equal to 3 right after the insertion, the newly joined replica cannot ACK
its own insertion into the `_cluster` space — if one out of original 3
nodes fails, then reconfiguration will fail.

Generally speaking, it will be impossible to join a new replica to the
cluster, if a quorum, which includes the newly added replica (which cannot
ACK), cannot be gathered.

To solve these problems, let's update the quorum in the `on_commit`
trigger. This way we’ll be able to insert a node regardless of the current
configuration. This somewhat contradicts with the Raft specification, which
requires application of all configuration changes in the `on_replace`
trigger (i.e., as soon as they are persisted in the WAL, without quorum
confirmation), but still forbids several reconfigurations at the same time.

Closes #10087

NO_DOC=<no special documentation page devoted to cluster reconfiguration>

(cherry picked from commit 29d1c0fa)
737dc12a
History
user avatar 737dc12a
History
Name Last commit Last update
.github
FreeBSD/databases/tarantool
apk
asan
changelogs
cmake
debian
doc
extra
patches
perf
rpm
rump
src
static-build
test
third_party
tools
test-run @ dd000632
.editorconfig
.gdbinit
.gitattributes
.gitignore
.gitmodules
.luacheckrc
.pack.mk
.test.mk
AUTHORS
CMakeLists.txt
CONTRIBUTING.md
Doxyfile
Doxyfile.API.in
LICENSE
README.FreeBSD
README.MacOSX
README.OpenBSD
README.md
TODO
README.md

Tarantool

Actions Status Code Coverage OSS Fuzz Telegram GitHub Discussions Stack Overflow

Tarantool is an in-memory computing platform consisting of a database and an application server.

It is distributed under BSD 2-Clause terms.

Key features of the application server:

Key features of the database:

  • MessagePack data format and MessagePack based client-server protocol.
  • Two data engines: 100% in-memory with complete WAL-based persistence and an own implementation of LSM-tree, to use with large data sets.
  • Multiple index types: HASH, TREE, RTREE, BITSET.
  • Document oriented JSON path indexes.
  • Asynchronous master-master replication.
  • Synchronous quorum-based replication.
  • RAFT-based automatic leader election for the single-leader configuration.
  • Authentication and access control.
  • ANSI SQL, including views, joins, referential and check constraints.
  • Connectors for many programming languages.
  • The database is a C extension of the application server and can be turned off.

Supported platforms are Linux (x86_64, aarch64), Mac OS X (x86_64, M1), FreeBSD (x86_64).

Tarantool is ideal for data-enriched components of scalable Web architecture: queue servers, caches, stateful Web applications.

To download and install Tarantool as a binary package for your OS or using Docker, please see the download instructions.

To build Tarantool from source, see detailed instructions in the Tarantool documentation.

To find modules, connectors and tools for Tarantool, check out our Awesome Tarantool list.

Please report bugs to our issue tracker. We also warmly welcome your feedback on the discussions page and questions on Stack Overflow.

We accept contributions via pull requests. Check out our contributing guide.

Thank you for your interest in Tarantool!