Fixity Auditing in Islandora

[kstapelfeldt]

Title (Goal)	Fixity Auditing in Islandora
Primary Actor	Collection Manager
Scope	Preservation
Story	As a collection manager I wish to be able to perform fixity audits on objects in the repository and see reports pertaining to the health of objects in my repository.

[rosiel]

Is this the same as "As a collection manager, I wish to generate checksums for the media in my repository."?

[kstapelfeldt]

They could be added together, but generating checksums would be discrete from checking checksums and seeing reports about it, I think?

[rosiel]

Possibly. Do you think there's a use case for "collection managers" to generate checksums that nothing happens with?

This was from:

• Feature Parity, line 84, phrased as "Enables checksums on datastreams".
• ICG Report, "Include and document digital preservation features in Islandora."

It sounds like this is part of a larger, "Preservation" interest/suite of features. I wonder if it'd be worth a dedicated interest group (or, "Initiative" which seem to be a bit different from our version of IGs)?

[mjordan]

Fixity auditing it verifying a file's checksum periodically over time to ensure that the file has not changed inadvertently. So generating checksums is not the same thing as fixity auditing, it's a precondition for fixity auditing.

[kstapelfeldt]

What Mark said. I think the use case would be that I have something else that checks the checksums. They could go together.

I think an interest group could be a great idea, but we know about immediate gaps in the toolset, so I think we can proceed to build things like checksum generation, and fixity checking and reporting (and ostensibly, a lot already exists in @mjordan work that could be embraced more officially)

[mjordan]

I am happy to contribute Riprap and Islandora Riprap Integration to the IF (and remain its maintainer). It currently generates checksums, performs the periodic verification of checksums, and reports checksum mismatches found during that periodic verification.

FYI Islandora already requires the Filehash contrib module, which generates checksums. AFAIK it doesn't do anything specific with them other than store them in a db table.

[mjordan]

@kstapelfeldt we have a Preservation IG but it has been inactive for some time. Could revive it.

[kstapelfeldt]

I think that could be a good longer-term goal for guiding preservation work in Islandora, so long as there is sufficient interest. For me, RipRap would be an excellent addition to the official repositories of Islandora.

[DonRichards]

@mjordan Is it possible to run RipRap without Fedora? Like storing a hash into a field or something 🤷

[kstapelfeldt]

I'm not the tagged person, but FWIW we're running RipRap without Fedora. It's definitely possible and Mjordan wrote this so it will work with Drupal.

[mjordan]

@kstapelfeldt is correct. You need to use the PluginFetchDigestFromDrupal plugin instead of the PluginFetchDigestFromFedoraAPI plugin.

[rosiel]

The Drupal Filehash module is already part of islandora, and we could improve how we use it!

• available since filehash 2.0, we can keep an "original" checksum?
• integrate with a local drupal checksum checker? Could it be a view?
• Current behaviour of islandora derivatives "replaces" files on disk, which would cause checksum check errors!
• Discuss best practices around duplicate checking
• Discuss best practices around hashing algorithms
• Make our hardcoded sha-1 check not rely on by luck having the sha-1 algorithm selected.

[dannylamb]

Yep, i think i've turned the 'original checksum' thing on a site or two. We're totally already keeping hashes of everything in the DB on drupal side. There's a lot to expand on there. Exciting!

[rosiel]

(Tech call) DGI has also released a Fixity Check module: https://github.com/discoverygarden/dgi_fixity