Yoav Tock

BFT Block Puller: A connection source that is updated immediately after a config block is sent to gossip

1

Right now the connection source (set of orderers) is updated when a config block is committed. We should change that so that the connection source is updated after a config...

Evicting an orderer while it is down, adding it again, and restarting it, w/o system channel, creates endless loop

2

**The problem**: 1. A 3-node cluster is started, say o1, o2, o3. (Raft IDs 1,2,3). 2. Orderer o1 is shutdown. 3. Orderer o1 is evicted, that is, removed from the...

The channel participation "Remove" does not clear the old etcdraft WAL & snapshots directories

1

The channel participation "Remove" removes the ledger but does not clear the old etcdraft WAL & snapshots directories that were used by the consenter. For example, starting a node with...

Change-Id: I89596f5d10e636e499c36799472e593582847a9e #### Type of change - New feature #### Description Add a replication policy to the orderer.yaml that allows a BFT orderer to choose whethet it uses the "simple"...

BFT smartbft integration test with synch with "simple" replication policy

1

If not specified, the default cluster replication policy is "consensus". This is what currently happens in all smartbft integration tests (it is not specified). We want to check whether the...

- Run a 4 orderer smart bft cluster. - Do 5 TXs. - Shut down o4. - Do 5 more TXs. - Shutdown o1,o2,o3 - Prepare 3 orderer mocks with...

This style of controlling the mock will only cover one use case. We want to achieve the same goal with a more flexible API on the mock, not by setting...

Currently the bft synchronizer is probing the height of remote orderers and deciding upon the target height using a heuristic based on cluster size and F. It is possible to...