Backup failed — what to check

When a backup fails, go to Backup History and click on the failed run to see the full error log. The log shows exactly which step failed and the error message.

SSH connection failures

• 'Connection refused' — SSH is not running on that port, or a firewall is blocking it
• 'Authentication failed' — wrong password or SSH key, or key not in authorized_keys
• 'SSH host key has changed' — the server presented a different host key from the one pinned on its first connection. Nothing was sent to it. If you rebuilt or reinstalled the server, check the new fingerprint and accept it on the server's page; if you did not, treat it as something answering in the server's place
• 'Connection timed out' — wrong IP address, or network issue

Database backup failures

• pg_dump: 'command not found' — PostgreSQL client tools not installed on the server
• pg_dump: 'permission denied' — SSH user doesn't have access to the database
• mysqldump: 'Access denied' — wrong database credentials
• mysqldump: 'Unknown database' — database name typo in the job config
• mongodump is not installed — mongodb-database-tools is a separate package from the MongoDB server; install it on the machine the job runs on
• mongodump: 'Authentication failed' — often the authentication database rather than the password. If the user was created in admin, set the job's authentication database field to admin

Storage upload failures

• 'Access Denied' — IAM/API key doesn't have PutObject permission on the bucket
• 'NoSuchBucket' — bucket name typo or wrong region
• 'RequestTimeTooSkewed' — your server's clock is out of sync (run ntpdate or chronyc)
• Endpoint errors — wrong endpoint URL for non-AWS providers (R2, B2, Spaces, Wasabi, UpCloud)

Re-running a failed job

Fix the underlying issue, then either wait for the next scheduled run or open the job and click Run now. A fresh backup run will be created — failed runs are never re-attempted automatically.