HeartBeatAgents

Off-Site Backup Won't Connect or Save

View as Markdown

What each specific connection-test error means, from rejected credentials to a bucket in the wrong region.

Off-site replication saving stays locked until Test connection passes; here's what each specific failure it can report actually means.

"The storage provider rejected these credentials"

The access key ID or secret itself is wrong. If you recently changed providers, re-enter both: credentials from one S3-compatible provider are never valid for another.

"The credentials are valid but lack permission to write to this bucket"

Authentication succeeded; authorization didn't. The access key needs write permission on this specific bucket, not just a valid account.

"Bucket not found"

Check the bucket name for a typo, and confirm it actually exists in the account these credentials belong to; a bucket that exists in a different account under the same provider looks identical to a missing one from here.

"The bucket exists in a different region"

The region setting has to match where the bucket actually lives, not where you'd prefer it to be. Check the bucket's real region in your provider's own console and match it, along with the endpoint URL, here.

"Could not reach the storage endpoint"

A network-level problem: the endpoint URL is unreachable from wherever this instance runs. Confirm the endpoint URL itself is correct before assuming it's a firewall issue; a typo in the endpoint produces exactly this error.

"The configuration is incomplete"

Bucket, access key ID, and secret are all required together; the test won't run at all until every one of them is filled in.

Provider-specific region notes

  • Cloudflare R2 ignores region entirely; "auto" is correct.
  • MinIO usually doesn't need a region set.
  • Any other S3-compatible provider: use whatever region matches your endpoint URL.