Serge Petrenko

core: fix crashes after altering trigger list while it is run

3

This patch fixes a number of issues with trigger_clear() while the trigger list is being run: 1) clearing the next-to-be-run trigger doesn't prevent it from being run 2) clearing the...

`before_replace` triggers inconsistency when updating a primary key

1

Tarantool version: 2.11.0-entrypoint-358-g9965e3fe4 (current master at the time of writing) I've noticed some strange behaviour with our before_replace triggers: 1. A usual insert of a duplicate key to a unique...

Now we enable Raft before a recovered node finishes syncing with others. This means that such a node may vote for anyone, even some node whose vclock is not the...

Before the patch IS_PRINTABLE macros was used to determine if given character is printable or not. This macros did not take into account characters encoded with 4 bytes. After the...

Linearizable read

11

Linearizability is a property of operations when operation performed on any node sees all the operations performed earlier on any other node of the cluster. More strictly speaking, it's a...

#6707 introduces linearizable transaction isolation level, but transaction isolation levels aren't supported by Vinyl engine at the moment (see #5522). Enable linearizable reads in vinyl once #5522 is fixed

Mixing asynchronous and synchronous replication leads to split-brain when a new synchronous transaction queue owner doesn't have all the async transactions of the former owner. The problem is that system...

There's no guarantee a PROMOTE entry from an old leader is present on the new one.

2

After becoming the Raft leader, the node writes PROMOTE in a separate worker fiber. Besides, the PROMOTE entry is replicated asynchronously. It's theoretically possible that a new leader is elected...

**Bug description** After upgrading the cluster from tarantool version 2.10.4 to 2.11.2 one of the replicasets had the following issue right on the first leader promotion: ``` main/145/applier/ I> remote...

[backport 3.1] relay: fix timeout on large xlog subscribe start

1

When a replica subscribes, it might in the beginning try to position its reader cursor to the end of a large xlog file. Positioning inside of this file can take...