You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, expirationd does not support iteration over a (secondary) index (since the expriration check callback must be applied to each tuple from a space). This creates a situation where a full scan is applied instead of selecting tuples from some start index value.
TODO:
Investigate how indices are used in the current expirationd implementation (check how the primary index is used, is it possible to iterate over a complex primary index, simple or complex secondary index)
Propose a RFC for changing the API to support iterating over a specified index or a field (the minimal satisfying index must be chosen in this case if it exists)
Check how the module is initialized: is it possible to reload the configuration and any created fibers or the internal state at runtime, thus supporting the "hot reload" feature
Propose a RFC for changes to support the "hot reload" feature
The text was updated successfully, but these errors were encountered:
We don't need this, since there is indexpiration already.
These modules have strongly different approaches: by index and by callback.
Mixing them up in one module will lead to many misuses.
I see no reason to implement it in expirationd. Different modules cause the user to choose between patterns: either full-scan with the arbitrary condition inside user function or very precise index scan by index with a timestamp.
Currently, expirationd does not support iteration over a (secondary) index (since the expriration check callback must be applied to each tuple from a space). This creates a situation where a full scan is applied instead of selecting tuples from some start index value.
TODO:
The text was updated successfully, but these errors were encountered: