Emergency / DR Backup Restoration Procedures
Not sure how to label or template this issue, my apologies. We are currently evaluating Stackgres and so far everything is very impressive!
After evaluating the backup / restore capabilities we are concerned about Emergency / Recovery situations where by the Stackgres installation itself becomes inoperable. Such situations could include:
- Breakage / corruption of the Stackgres installation itself, either through user error or a future bug
- Loss of the underlying kubernetes cluster itself, whereby recovery is not possible, and the original Stackgres installation cannot be restored.
It seems Stackgres uses well known community tooling WAL-G to perform backups. In our testing we've configured backups to land in an Azure Blob Storage Account and it works well. However I cannot find any documentation on how to manually restore the persisted WAL-G backups to another pg instance or a new Stackgres instance cluster.
I hope to discuss the possibility of the following:
- The ability for a Stackgres instance to
discoverbackups performed by some other Stackgres instance, by pointing Stackgres to an existing backup location in Azure Blob, S3 etc. This would enable a Emergency recovery scenario where a new Stackgres instance is provisioned, and backups from the old Stackgres instance can be restored unto new Stackgres PG clusters during creation time. - Documented "break-glasss" procedures on how to use the WAL-G tooling directly to restore backups created by Stackgres, into pg instances provisioned by a new Stackgres instance or pg instances provisioned classically.
Thanks for any info. Even with these concerns I can say the quality and UX of evaluating Stackgres so far has been fantastic!