I loaded an empty database like this:

time bin/init_osm3s.sh /home/ajtown/data/north-yorkshire_2026-02-12T21:23:29Z.osm.bz2 "db/" "./" --meta=no

Set replication_id to replicate_id to 6984689 and ran the server via db/startup.sh

I can query the database from the command line with "bin/osm3s_query" and

<query type="nwr"><bbox-query n="54.0" s="53.98" w="-1.09" e="0.01"/><has-kv k="amenity" v="pub"/></query><print/>

returns some data as expected

The downloaded diffs are as expected:

/opt/op/diff/006: (1.3 GiB available)
drwxrwxr-x 5 op op  4096 Feb 13 20:28 .
drwxrwxr-x 3 op op  4096 Feb 13 23:05 ..
drwxrwxr-x 2 op op 28672 Feb 13 20:27 984
drwxrwxr-x 2 op op 69632 Feb 13 20:28 985
drwxrwxr-x 2 op op 20480 Feb 13 23:05 986


/opt/op/diff/006: (1.3 GiB available)
drwxrwxr-x 5 op op  4096 Feb 13 20:28 .
drwxrwxr-x 3 op op  4096 Feb 13 23:05 ..
drwxrwxr-x 2 op op 28672 Feb 13 20:27 984
drwxrwxr-x 2 op op 69632 Feb 13 20:28 985
drwxrwxr-x 2 op op 20480 Feb 13 23:05 986

Only one of those was over 1M; the original extract was 56MB. However, here are the files in the db directory:

op@h22:~/db$ du -BG * | sort -n -r | head
23G nodes.map
8G ways.map
1G way_tags_local.bin.shadow
1G way_tags_local.bin.idx.shadow
1G way_tags_local.bin.idx
1G way_tags_local.bin
1G way_tags_local_attic.bin.shadow
1G way_tags_local_attic.bin.idx.shadow
1G way_tags_local_attic.bin.idx
1G way_tags_local_attic.bin

This is surely either implausibly large or extremely inefficient?

Is there any documentation anywhere that describes what data format nodes.map is in, how I can see what is in it or how I can control how much space is used?