What happens with files in /repo/blob?

This is my understanding so far:

Files in the /binary folder are never deleted? If I upload two identical images they get the same hash and there is no need to overwrite or save the second one? So this folder get mostly reads and not so many writes?

But what about /node? Is it the same behavior? Since all versions of a content is available I guess each time a content changes it gets a new file?

The reason I ask is I took a backup and then did a restore with the snapshot tool. What happend was my user content had changed between snapshot and backup. Looked like the lastLogin timestamp had changed and content studio crashed because my user is probably the owner of everything and it coundln’t find the new blob file for my user

Does every login result in a new document in the /node folder just because a timestamp changes?

You are right on how the blobstore works, it will write new files for each new binary or version. This is a bit greedy in terms of disk, but it is also fast.

When it comes to your issue, it is controlled by the Idprovider if it changes anything on login - looks like it does in your case.

Even if every login creates a new version of the user node, there are no links to a specific version of a user (or other data nodes for that case). Users are actually referenced with a principalkey in the format: user:<userstore>:<username>.

So, I’m not sure what actually causes your problem? Is it likely that the username (loginid) changed?

Also, restoring with the XP snapshot tool will never delete or create any new blobs.

Ok, I understand. So hypotheticaly, If you have 200.000 users logging into your site every day you get 200.000 new versions?

Hypotetically, yes…

How does the vacuum feature of the toolbox work? What files will be deleted?

All repo-files on disk that are not connected to any node will be deleted.
Old versions of a node that exists will NOT be deleted.

1 Like

This topic was automatically closed after 2 days. New replies are no longer allowed.