I'm continuing to investigate the strange file cleaner behavior i'm seeing (mentioned in this thread
937065 and wanted to make sure i understand correctly how it works.
when it cleans a file, its purely moving the storage for a given node to the new file, its not in anyway mutating the tree structure itself? and given no application changes to the db, you'd expect the new files generated to have 100% utilization? What i appear to be seeing is that the cleaner is generating new files, then shortly after is cleaning those files it just generated, and it appears to just go and go.
I have got a separate copy of the db that i'm experimenting with, running DbSpace shows that most of the files have utilizations in the 40 - 60% range, but strangely even the recently created files have 40-70% utilization. I'm planning to write a new host process that will ensure that no changes to the data are being made by the application, and see what the file cleaner does. We're using 3.3.82, although the db was originally created by an earlier version, possibly 3.3.24 (I need to double check).
If anyone can shed any light on possible issues or things to investigate, I'd appriciate it.
Thanks
Simon