Hedge
Search…
Archiving

Archiving

Copying data to tape is called Archiving, as its goal is to be an independent copy of data that doesn't rely on other backups to be retrievable. Archiving with Canister is simple:
First, select the disk or folder you want to archive by either dragging in a folder from Finder, or by browsing to a Source Folder. Then continue, and mount a tape. Next, you'll confirm the transfer(s).
Often, files that appear to consecutively and part of the same folder, are not actually stored in consecutive order. When working with LTO, it's important that files are lined up and written in consecutive order to prevent shoe-shining upon retrieval. Therefore, it's best practice to copy all files into a single folder, and archive that folder, so that Canister can make sure the files are written to tape in the best order possible.
In a future update, we'll add queueing and source collections to Canister.

Duplicate Detection

Canister automatically applies Duplicate Detection to all transfers. When a file is identical in name, size, and modification date, it is skipped. When a file has been changed, the already stored file will be renamed by appending its modification date, and hidden. That means all previous versions of a file will always be available for retrieval at a later time.

Incremental Backups

As Canister comes with Duplicate Detection, it will detect what is new or has changed and only copy those files and folders to tape. Keep in mind that due to the linear nature of LTO, this will cause your data to be stored in different sections of the tape, and thus will impact retrieval time as the tape head has to make a lot more movements to access a folder's contents across a tape.

Simultaneous Backups

If you have multiple tape drives, and your license supports it, Canister can write to multiple tapes simultaneously. If the tapes used are of equal speed, it will be as fast as writing a single tape. It does help to use a faster-than-tape source, to speed up verification.

Illegal Characters

Many legacy systems use a deny-list of characters, severely complicating many workflows. To solve that without having to rely on a database and thus introducing vendor lock-in, Canister utilizes percent-encoding for all illegal and unsupported characters. You cannot only use any of the forbidden characters / * ? < > " | \ : but also use any internationalization in all file and folder names. Canister will know when it needs to replace them, and when not to.
Upon retrieval, Canister automatically converts percent-encoded characters back to Unicode. If you make use of illegal characters heavily and plan to retrieve a Canister-made LTO with a different mechanism or app, it's easy to use a percent-decoding script after doing the retrieval.

Verification

By default, each archive is fully verified by reading back the files after been having transferred. Verification can be skipped at any time, by clicking the x button next to the progress bar.
Canister uses XXH64BE. For legacy workflows, it's also possible to use MD5 instead, or even to disable verification altogether:

Compression

LTO drives offer built-in hardware compression. Compression is a property of a tape and is set to use compression when formatting the tape. After the property has been set, it cannot be removed without again erasing the tape. As the compression has zero to no overhead, there's no reason to disable it.
As the compression happens in-device, there's is no feedback on the process, or on how much data is actually saved. Although touted as a big feature for LTO, don't expect any wonders from it for media archival. LTO compression only works well for text files and not at all for video or photon material, so never expect to be able to save more data to a tape than the bare capacity.

Encryption

Canister currently does not support hardware-based LTO encryption.