Proto commits in liftbridge-io/liftbridge

These 80 commits are when the Protocol Buffers files have changed:

2022-03-16

Add expired field to LeaveConsumerGroupOp This indicates if the consumer was removed because they timed out.

The documentation is generated from this commit.

2022-01-20

Implement group coordinator failover

2021-12-31

Combine assignmentEpoch and coordinatorEpoch Use a single consumer group epoch.

2021-12-07

Handle consumer groups on subscribe

2021-11-18

Consumer groups WIP

2021-04-25

rename encryption variables

2021-03-20

Add config for EncryptionDataAtRest - Add default config - Add config overwrite

2020-12-12

Add OptimisticConcurrency as a config option on CreateStream

2020-10-18

Fix stream creation time not being set once by the leader

2020-10-12

Add readonly status to metadata

2020-10-03

Support overriding MinISR on CreateStream Addresses #269.

2020-09-29

Merge branch 'master' of github.com:liftbridge-io/liftbridge into consumer_tracking_api

2020-09-15

Cursor Manager WIP

2020-09-14

Add read-only flag to streams/partitions Readonly streams: simplify and don't block if readonly Fix tests Add comments Add tests Fix a comment

2020-09-07

Add option to disable partition auto pausing if subscribers > 0

2020-09-04

Implement SetStreamConfig endpoint

2020-07-20

Implement automatic partition pausing

2020-06-26

Refactor custom stream config Also ensure custom config survives Raft snapshot/restore.

2020-06-21

Support nullabletype in custom stream config

2020-06-13

Use nested message for CustomStreamConfiguration

2020-05-31

Add CompactEnabled into stream request parsing

Use int64 as configuration type for stream duration configs

2020-05-23

WIP: use string for duration stream config

2020-05-21

Renaming CustomStreamConfig

2020-05-09

Drop support for Compact

2020-05-08

Use gogo/protobuf for generating duration

2020-05-03

WIP: Add custom streams config to proto

2020-04-13

Merge branch 'master' of github.com:liftbridge-io/liftbridge into improve_activity_stream

2020-04-10

Send leader epoch on replication requests Send the leader epoch on replication requests. When a partition leader receives a replication request, check that the leader epoch is the current leader's epoch. If not, drop the request. This fixes an edge case where a leader is partitioned from the rest of the ISR and a new partition leader is elected. Previously, both the old leader and new leader would receive the replication requests. Followers would reject responses from the old leader, but the old leader would still commit messages due to the replication requests sent by followers and stay in a zombie state.

2020-04-09

Add other events to activity and add tests

2020-04-08

Refactor raft operations to be more consistent

Refactor raft operations

2020-04-06

Activity stream manager WIP

2020-03-31

Rename proto package to avoid protobuf collision Rename the internal proto package to protocol to avoid protobuf collision with the client proto package.

2020-03-26

Refactor stream pausing Also update go-liftbridge and liftbridge-api deps.

2020-03-24

Add partition pausing and resume all at once options

2020-03-23

Add partition pausing and auto resume on publish

2020-03-11

Add a delete stream implementation

2019-10-16

Add unit tests around LEO notifications And other general code cleanup.

2019-10-15

Fix idle follower ack delay Fix issue #102 by having the leader notify followers when there is new data available after the follower has caught up with the leader's log end offset. This also introduces a new config `replica.max.idle.fetch.wait` which controls the max amount of time an idle follower will wait before making a replication request. The primary purpose of this is to ensure timely check-ins with the leader for health checking.

2019-08-28

Clean up metadata api

2019-08-22

Remove TotalPartitions field from partition

Rename name to stream

2019-08-16

Partitioning WIP

2019-07-25

Implement partitioning

2019-07-17

Alter replication protocol to use leader epoch Currently, when a replica starts, it truncates its log up to the high watermark (HW) to remove any potentially uncommitted messages. There are a couple edge cases with this method of truncating the log that could result in data loss or replica divergence. This solves the issue by using a leader epoch rather than the high watermark for truncation. Fixes #38

2019-07-02

support github.com/nats-io/nats-server/v2 and github.com/nats-io/nats.go 1. support github.com/nats-io/nats-server/v2 and github.com/nats-io/nats.go 2. move bridge-io/nats-on-a-log inside repo

2019-06-28

Update go-liftbridge dependency

2019-04-16

Fix module name

2019-04-13

Refactor to Go modules

2018-08-30

Update go-liftbridge again

Update go-liftbridge

2018-08-28

Update go-liftbridge

2018-08-10

Update go-liftbridge dep

2018-08-08

Update deps

2018-07-23

Update liftbridge-grpc dep

Update deps

2018-07-17

Use correct go-liftbridge dependency

Update import paths

2018-07-03

Fix stuff using new subscribe api

Fix publish race condition Also wire up correlation id properly.

Vendor dependencies

2018-07-02

Use correct path for grpc definitions

2018-06-29

Rename ConsumerGroup to Group

2018-06-24

Add replication test and fix create stream race

2018-06-09

Make server id durable

2018-06-01

Refactor leader failure detector

2018-05-31

Reimplement replication to improve performance

2018-05-22

Implement fetch metadata api

2018-05-17

Rename jetbridge to liftbridge

2018-05-15

Fix stream recovery from raft log

Lots of bug fixes

2018-05-14

Implement leader failure detector Also implement leader failovers and expanding ISR.

2018-05-10

Implement high watermark and log truncation

2018-05-04

Implement request propagation and ISR shrinking

Start sketching out replication

2018-05-03

Rename participant to replica

2018-05-02

Refactor streams and implement snapshot/restore

2018-05-01

Implement log

2018-04-27

Initial commit