When collecting a cluster backup in a sharded cluster it's important not to do so in the middle of rebalancing. Otherwise the data might be lost (shard A backed up after bucket move to shard B, shard B backed up before bucket arrival) or duplicated.
Also when restoring from backups with xlogs to a particular point in time, the same may happen.
We need to export information about the times when rebalancing happened in cluster so that the user is able to rule out which points in time are safe to restore to and which aren't.
JIRA issue: https://jira.vk.team/browse/TNTP-7360
When collecting a cluster backup in a sharded cluster it's important not to do so in the middle of rebalancing. Otherwise the data might be lost (shard A backed up after bucket move to shard B, shard B backed up before bucket arrival) or duplicated.
Also when restoring from backups with xlogs to a particular point in time, the same may happen.
We need to export information about the times when rebalancing happened in cluster so that the user is able to rule out which points in time are safe to restore to and which aren't.
JIRA issue: https://jira.vk.team/browse/TNTP-7360