Details
-
Task
-
Resolution: Unresolved
-
Major
-
7.6.2
-
0
Description
What is the problem?
In MB-61550 query was changed so that it returns an error from the backup endpoint if the migration from pre-7.6/_system scope metadata storage has not happened yet. This means that there might be a brief period when a >=7.6 first starts where backups fail with
Error backing up cluster: internal server error executing 'GET' request to '/api/v1/backup': {"_level":"exception","caller":"admin_accounting_endpoint:1011","cause":{"possible_reasons":["Migration active."]},"code":7300,"key":"service.backup_not_possible","message":"Backup not possible currently."}
|
We seem to hit this pretty reliably in examinador which makes sense as we start a cluster, load very few items and then immediately take a backup.
What is the solution?
As part of the cluster readiness check we should wait until the query metadata migration completes.
Attachments
Issue Links
- relates to
-
MB-61550 Backup bucket endpoint errors if cluster is in mixed-mode due to _system scope
- Resolved