Deletion: Handle the case of pruned objects. #5885

Closed
lambadalambda wants to merge 0 commits from gitlab-mr-iid-2507 into develop

Closes #1204

I'd recommend removing pruning completely, it makes ensuring a consistent database state much harder, for little gain. We should use other mechanisms to prevent extensive space usage over time.

Closes #1204 I'd recommend removing pruning completely, it makes ensuring a consistent database state much harder, for little gain. We should use other mechanisms to prevent extensive space usage over time.
Author
Owner

Also, while this should probably be merged into stable, this part changed so much it's hard to port. Don't really know what to do about this except writing a separate fix for stable.

Also, while this should probably be merged into stable, this part changed so much it's hard to port. Don't really know what to do about this except writing a separate fix for stable.
Member

"little gain" is much arguable. However I get the concern, I think we should just support backdating activities, which would allow us to remove the creates with objects as if they were never there

"little gain" is much arguable. However I get the concern, I think we should just support backdating activities, which would allow us to remove the creates with objects as if they were never there
Author
Owner

Yeah, we should remove the create activities with it, then it wouldn't be much of an issue.

I absolutely see the concern for dbs getting too large, but pruning objects is not the right approach, especially if it complicates the database state this much. Pruning can't even save much space, it only deletes objects, which are only a small part of overall remote data.

If some remote users posts something, we get a create activity and an object. Now 20 remote people we follow click like on that, we now have 20 more activities. Then pruning happens some time and instead of 22 database objects of similar size we now have 21. That's what I mean by little gain. Reducing disk usage is necessary, but pruning (in the current way) is not the right approach.

We also have two different pruning mechanisms, one is Object.prune (which as far as I can tell isn't used at all) and the one in the mix task. I think a task that would just delete all remote activities older than a cutoff date would make more sense.

Yeah, we should remove the create activities with it, then it wouldn't be much of an issue. I absolutely see the concern for dbs getting too large, but pruning objects is not the right approach, especially if it complicates the database state this much. Pruning can't even save much space, it only deletes objects, which are only a small part of overall remote data. If some remote users posts something, we get a create activity and an object. Now 20 remote people we follow click like on that, we now have 20 more activities. Then pruning happens some time and instead of 22 database objects of similar size we now have 21. That's what I mean by little gain. Reducing disk usage is necessary, but pruning (in the current way) is not the right approach. We also have two different pruning mechanisms, one is `Object.prune` (which as far as I can tell isn't used at all) and the one in the mix task. I think a task that would just delete all remote activities older than a cutoff date would make more sense.

Pull request closed

Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
2 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
pleroma/pleroma!5885
No description provided.