Closed
Description
Currently the gitdb will memory map pack indices, and loose objects for reading. If there are too many mapped files, which could happen on large databases, one has to gracefully unload existing memory maps and reload them later.
This could be as easy as deleting the data caches, which are memory maps most of the time.
This task could be done by the PackedDB, and additionally by the git-db which handles multiple object databases.
Metadata
Metadata
Assignees
Type
Projects
Relationships
Development
No branches or pull requests
Activity
Byron commentedon Jan 7, 2015
I think this now works, see #60 .
The memory manager is able to unload existing maps if it hits a resource limit (soft or hard).