The idea is that using a secondary (e.g. in a multiplexer configuration) objstorage would allow us to be able to keep a copy of the few content that are colliding on sha1.
It is not a proper long term solution.
Description
Description
Revisions and Commits
Revisions and Commits
| rDDATASET Datasets | |||
| D8008 | rDDATASETe31bdb26a827 Set object id when calling objstorage.add | ||
| rDOBJS Object storage | |||
| D8756 | rDOBJSdf4be2d87c30 azure: Add tests based on Azurite in addition to mocks | ||
| D8029 | rDOBJS667cb87b9367 Start introducing composite ObjId in the interface | ||
Event Timeline
Comment Actions
do you have in mind to make the actual hash used as primary key in an objstorage a configuration of said storage instance? e.g. create a pathslicer or s3 objstorage using sha256 is just a matter of configuration of the objstorage?
Comment Actions
Also, is the idea is to make any swh objstorage able to be queried for a content using any supported hash? or will the only query API require a multihash object?
Comment Actions
Possibly relevant for the Azure storage: https://learn.microsoft.com/en-us/rest/api/storageservices/find-blobs-by-tags