View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0000812||LDMud 3.5||Runtime||public||2012-11-11 23:00||2012-12-04 17:01|
|Target Version||3.5.0||Fixed in Version||3.5.0|
|Summary||0000812: Data cleanup behaviour should be improved|
|Description||Currently the data cleanup behaviour is somehow flawed.|
The number of objects cleaned per backend cycle is dependent on the TIME_TO_CLEANUP (i.e. the time before calling cleanup()). This dependency does not make much sense. If you want to have calls to cleanup() only for objects which were not used for a long time, you choose a large TIME_TO_CLEANUP. But because the number of objects data-cleaned is 2*number_of_objects / time_to_cleanup, this leads to a very low number of objects cleaned per cycle.
2) The first X objects of the processed objects in a given backend cycle are data-cleaned. If all objects are processed in a cycle, always the same X objects are data-cleaned and the other never. (Unless the object list changes.)
|Tags||No tags attached.|
I changed the default for the data cleanup time to 3600 s and made it configurable with configure_driver(). Additionally a bug was fixed during object creation (the time was initialized with the genereal cleanup time).
In 3.3 the default was changed to 10800 s (longer, because not configurable at run-time) and the mentioned bug during object creation was fixed.
|2012-11-11 23:00||zesstra||New Issue|
|2012-12-04 14:11||zesstra||Assigned To||=> zesstra|
|2012-12-04 14:11||zesstra||Status||new => assigned|
|2012-12-04 17:01||zesstra||Note Added: 0002155|
|2012-12-04 17:01||zesstra||Status||assigned => resolved|
|2012-12-04 17:01||zesstra||Fixed in Version||=> 3.5.0|
|2012-12-04 17:01||zesstra||Resolution||open => fixed|