Initial commit to add head_only mode to leveled. This allows leveled to receive batches of object changes, but where those objects exist only in the Penciller's Ledger (once they have been persisted within the Ledger).
The aim is to reduce significantly the cost of compaction. Also, the objects ar enot directly accessible (they can only be accessed through folds). Again this makes life easier during merging in the LSM trees (as no bloom filters have to be created).
If wast retention period is undefined, then it should be ignored - and no waste retained (rather than retaining waste for 24 hours as at present).
This wasn't working anyway - as reopen reader didn't get the cdb options (which didn't have the waste path on anyway) - so waste would not eb retained if the file had been opened after a stop/start.
Compression can be switched between LZ4 and zlib (native).
The setting to determine if compression should happen on receipt is now a macro definition in leveled_codec.
Introduce a dedicated module for all the different fold types. Also simplify the list of folders by deprecating those folds that should eb achieveable by fold_heads/fold_objects type folds but with smarter functions.
Makes sure that the fold functiosn also have better spec coverage, and are dialyzer checked.
Cmpaction is overly aggressive. It is a lot of work to compact a run of
files for just 20% reduction in disk space, when disk space for the
Journal (i.e. low IOPS disk space should be relatively inexpensive).
Require at least a 40% reduction for a compaction job.
When running a load of mainly 2i queries, there is a huge cost in the
previous snapshot code. The time taken to create a clone of the
Penciller (duplicating all the LoopState) varied between 1 and 200ms
depedning on the size of the LoopState.
For 2i queries, most of that LoopState was then being thrown away after
running the query against the levelzero_cache. This was taking < 1ms on
average. It would be better to avoid the o(100)ms of CPU burning and
block for o(1)ms - so th eorder of events have been changed ot filter
first so only the small part of the LoopState actually required is
copied to the clone.
Move to using the DJ Bernstein Magic Hash consistently, and trying to
make sure we only hash once for each operation (as the hash is more
expensive than phash2).
The improved lookup time for missing keys should allow for the L0 index
to be removed, and hence speed up the completion time for push_mem
operations.
It is expected there will be a second stage of creating a tinybloom as
part of the SFT creation process, and then adding that tinybloom to the
manifest. This will then reduce the message passing required for a GET
not in the cache or higher levels
