Help understanding log - contents vs blobs?

General Topics
1 
16:42:14.713 GC found 33894 unused contents (18.2 GB)
16:42:14.713 GC found 0 unused contents that are too recent to delete (0 B)
16:42:14.713 GC found 244346 in-use contents (711.1 GB)
16:42:14.713 GC found 17 in-use system-contents (9 KB)
16:42:14.718 Previous content rewrite has not been finalized yet, waiting until the next blob deletion.
16:42:14.718 Found safe time to drop indexes: 2025-05-18 13:19:38.7184843 +0100 BST
16:42:14.718 Dropping contents deleted before 2025-05-18 13:19:38.7184843 +0100 BST
16:42:14.807 Looking for unreferenced blobs...
16:43:13.210 deleted 100 unreferenced blobs (2.3 GB)
16:43:54.832 deleted 200 unreferenced blobs (4.6 GB)
16:44:24.093 deleted 300 unreferenced blobs (7 GB)
16:44:46.361 deleted 400 unreferenced blobs (9.4 GB)
16:45:03.000 deleted 500 unreferenced blobs (11.6 GB)
16:45:03.454 Found 521 blobs to delete (12.1 GB)
16:45:03.454 Deleted total 521 unreferenced blobs (12.1 GB)
16:45:03.460 Extending object lock retention-period is disabled.
16:45:03.460 Compacting an eligible uncompacted epoch...
16:45:04.898 Cleaning up no-longer-needed epoch markers...
16:45:04.900 Attempting to compact a range of epoch indexes ...
16:45:04.901 Cleaning up unneeded epoch markers...
16:45:04.906 Cleaning up old index blobs which have already been compacted...
16:45:04.965 Keeping 2 logs of total size 5.8 MB
16:45:04.965 Cleaned up 0 logs.
16:45:04.966 Finished full maintenance.

So, 18.2 GB unused contents but only 12.1 GB are deleted. Contents vs blobs? Could someone help clarify?

2

“contents” refers to blocks of well… content, while blobs are containing one or more content objects. content blocks is what makes deduplication work and one content block will be at least referenced by one file in your repo - hopefully by more than one, which is, what dedup aims at.

Afair, the size of a blob is up to 20 MB, but can vary and blobs are used to keep the number of objects in a repo at bay. This is also the reason that maintenance will always to work on blobs, when contents get aged out of the repo.

In regards to unused contents vs. unreferenced blobs: unused contents will only be actually removed after its retention period hass passed, it will linger behind in the repo until that time.

3

Hi. Thanks, that does shed some light on it all - even if I still have a rabbit warren of questions on how it all works!

Cheers.