Do not use write-cache as a read cache: always remove objects from the
WC, not only if an object hasn't been used for some time (LRU cache is
dropped). Use object size (in bytes) as a metric of used space, not an
approximate (and too inaccurate) maximum stored objects number.
Signed-off-by: Pavel Karpy <p.karpy@yadro.com>
It is not used for testing, its absence does not break build. The only one
implementation is placed in the same package.
Signed-off-by: Pavel Karpy <p.karpy@yadro.com>
Do not use WC's internals in the initialization routines without mode
protection. WC should be able to change its mode even if the initialization
is not finished yet.
Signed-off-by: Pavel Karpy <p.karpy@yadro.com>
In case we have many small objects in the write-cache, `indices` should
not be reused between iterations.
Signed-off-by: Evgenii Stratonikov <e.stratonikov@yadro.com>
If an object has not been marked for removal by the GC in the current epoch
yet but has already expired, respond with `ErrObjectNotFound` api status.
Also, optimize shard iteration: a node must stop any iteration if the object
is found but gonna be removed soon.
All the checks are performed by the Metabase.
Signed-off-by: Pavel Karpy <carpawell@nspcc.ru>
