S3 Bucket Cleanup

I’m playing with oCIS 4.0.2, deployed via Helm Chart.

Apart from some issue I already reported and some points which are on my TODO to recheck. Everything seems to be running fine so far.

I’m utilizing Ceph Radosgw as storagebackend. Performance is great so far, but I’ve encountered two more-or-less pressing topics for me:

  1. I like the background sync towards S3, although it would great to have some indication in the UI on this. Right now I get an error if I try to download the same file right away.
  2. As I’m currently in testing-mode. I’m constantly adding and deleting files. I noticed my S3-Bucket is filling with orphans. Files are deleted from the trash-bin in the webUI. I’ve enabled both cleanup jobs in the Helm Chart and set DeleteBefore options to 60s. Data remains in the bucket. Is there some kind of additional house-keeping running in the background? I would also not rule-out configuration issues on my side.


Did some further testing on #2:
I have no idea what happened, but as of now I cannot reproduce it.

I have some orphans in that bucket, which are totally fine. Outdated from former testing where I was just to lazy to do a cleanup. But there are a bunch of objects I cannot correlate to a killed testing cycle.

So rephrasing my question: Is there any way for me to do a correlation between whats in the metadata index and whats stored in the bucket?

Unfortunately this is currently not easily possible. They can be correlated by the blobid which is a metadata field of the item. I described how to get them in an old ticket: Issue Recovering Deleted Files in oCIS · Issue #6399 · owncloud/ocis · GitHub
But there is no way to get a list of them through ocis. One would need to use a shell script or similar to get them.

But I have good news regarding question #1: We are already working on this. It should be available with the next major release.

Doesn’t matter if complicated or not, at least there is a way to do that :slight_smile:
Is oCIS constantly “scanning” the msgpack every time you access a file? If not, is this information stored in e.g. NATS or Redis?

The indicator is amazing news.

Yes, ocis is always checking the messagepack every time a file is accessed. That is the way metadata is stored in ocis.
NATS or redis are only used for caching with a low ttl. So looking for data there is not recommended.