Support

Offset 19 is the end of the archive header (the bit that identifies the file as a backup archive) and the beginning of the actual backup data. If you get an error there it means that your backup archive was damaged or truncated.

Alternatively, it could mean that you don't have all of the backup archive parts. If the backup consisted of more than one files (e.g. .jpa, .j01, .j02, ...) you need to have all of these files present in the same directory for the extraction to work. They are not different backup archives. They are the integral parts of a single backup archive. Think of it as a magazine page torn into many pieces. Each piece is useless. You need to put together all pieces to read the content of the page. In our case, if you only have the .jpa file it makes sense you get this error message: .jpa is the last part therefore offset 19 most definitely does not contain the beginning of file data. This is what happens with the first part file of the backup archive, i.e. the .j01 file. So first make sure you have all the backup archive parts in the same folder.

If you are sure you have all the backup archive parts and / or there was only one part (just a .jpa file) try enabling the "Ignore most errors" option in Kickstart. This uses some heuristics to try and find the file data in damaged archives. If that doesn't work the file is well beyond repair. In the latter case you should try and restore from an older backup.

There are ways to avoid being in this situation in the future.

We recommend using the "Archive integrity check" option when taking backups and testing your backups regularly. You can even automate the test restoration of backups using Akeeba UNiTE. As the old adage goes, you either have tested backups or no backups. This applies to all backups taken with any software, not just site backups taken with Akeeba Backup.

Furthermore, you should transfer the backup archives either with FTP in Binary transfer mode (not Auto, not ASCII) or SFTP or have them automatically uploaded to remote storage. In the latter case you should check their integrity because remote storage providers don't guarantee 100% safety of your data and their APIs can fail during transfer in ways that don't report an error but don't result in a complete data transfer either. Don't use the browser to copy files, they are likely to be truncated.