-
Type: Task
-
Resolution: Done
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: Cloud Manager
-
Labels:None
As part of BRS-1553, we will start taking cluster snapshots regardless of whether the balancer was stopped. There will be additional terse warnings in the UI and we need docs to better explain the state of the clustershots as well as the risks of restoring them.
1. Clustershot taken with the balancer running: The agent timed out waiting for the balancer to stop or current migrations to complete. Snapshots were taken and the clustershot is restorable. Possible data loss or orphan data if the clustershot is restored. Individual shards are restorable as normal.
2. Clustershot taken with 1 or more shards/config servers unreachable. The agent couldn't reach one of the shards/config server to insert a synchronization oplog token. Clustershot is not restorable, but the shards that succeeded were snapshotted and are restorable.
See https://github.com/10gen/mms/pull/2184 and https://github.com/10gen/mms/pull/2262 for screenshots.
This is scheduled for Aug 19th release.