The file is sparse, which really is supported by all sane filesystems
nowadays I think. But for 32-bit systems I'm a bit afraid for the
ability to reliably get such a large contiguous mapping in process memory,
so there we take the 500M limit tested in knot-dns:
https://gitlab.nic.cz/knot/knot-dns/-/blob/v3.3.5/src/knot/conf/schema.c#L39
- docs: documentation is now separated into user and developer parts (!1514)
- daemon: ignore UDP requests from ports < 1024 (!1507)
- manager: increase startup timeout for processes (!1518)
+- local-data: increase default DB size to 2G on 64-bit platforms (!1518)
Bugfixes
--------
// FIXME: the file will be sparse, but we still need to choose its size somehow.
// Later we might improve it to auto-resize in case of running out of space.
// Caveat: mdb_env_set_mapsize() can only be called without transactions open.
- .maxsize = maxsize ? maxsize : 100 * 1024*(size_t)1024,
+ .maxsize = maxsize ? maxsize :
+ (sizeof(size_t) > 4 ? 2048 : 500) * 1024*(size_t)1024,
};
int ret = the_rules->api->open(&the_rules->db, &the_rules->stats, &opts, NULL);
/* No persistence - we always refill from config for now.