Skip to content

dr-autosync: canSync and hasMajority can be calculated wrong when placement rule config is complex #7201

@disksing

Description

@disksing

Bug Report

What did you do?

deploy a dr-autosync cluster, setup 3 voters in primary dc, 2 voters in dr dc. the tikv-servers in dr dc splits into 2 different logic zones and use placement rule to place 1 voter in each logic zone.

What did you expect to see?

kill all tikv-servers in a logic zone, the dr-autosync state should not switch to async_wait or async, as there is still 1 voter in dr zone.

What did you see instead?

the state is switched to async_wait and then async.

What version of PD are you using (pd-server -V)?

master / v6.5.5

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

Status

Closed

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions