Merge pull request #961 from lonvia/harmonize-cache-documentation

put different advise on cache size in sync

Author: pnorman

docs/osm2pgsql.1

@@ -112,9 +112,11 @@ imported into database columns and which tags get dropped. Defaults to /usr/shar
 \fB\-C\fR|\-\-cache num
 Only for slim mode: Use up to num many MB of RAM for caching nodes. Giving osm2pgsql sufficient cache
 to store all imported nodes typically greatly increases the speed of the import. Each cached node
-requires 8 bytes of cache, plus about 10% \- 30% overhead. For a current OSM full planet import with
-its ~ 3 billion nodes, a good value would be 27000 if you have enough RAM. If you don't have enough
-RAM, it is likely beneficial to give osm2pgsql close to the full available amount of RAM. Defaults to 800.
+requires 8 bytes of cache, plus about 10% \- 30% overhead. As a rule of thumb,
+give a bit more than the size of the import file in PBF format. If the RAM is not
+big enough, use about 75% of memory. Make sure to leave enough RAM for PostgreSQL.
+It needs at least the amount of `shared_buffers` given in its configuration.
+Defaults to 800.
 .TP
 \fB\  \fR\-\-cache\-strategy strategy
 There are a number of different modes in which osm2pgsql can organize its

docs/usage.md

@@ -25,11 +25,15 @@ use them.
 Performance is heavily influenced by other options, but there are some options
 that only impact performance.
 
-* ``--cache`` specifies how much memory to allocate for caching information. In
-  ``--slim`` mode, this is just node positions while in non-slim it has to
-  store information about ways and relations too. The maximum RAM it is useful
-  to set this to in slim mode is 8 bytes * number of nodes / efficiency, where
-  efficiency ranges from 50% on small imports to 80% for a planet.
+* ``--cache`` specifies how much memory in MB to allocate for caching information.
+  In ``--slim`` mode, this is just node positions while in non-slim it has to
+  store information about ways and relations too. The rule of thumb in slim mode
+  is as follows: use the size of the PBF file you are trying to import or about
+  75% of RAM, whatever is smaller. Make sure there is enough RAM left for
+  PostgreSQL. It needs at least the amount of `shared_buffers` given in its
+  configuration. You may also set ``--cache`` to 0 to disable node caching
+  completely. This makes only sense when a flat node file is given and there
+  is not enough RAM to fit most of the cache.
 
 * ``--number-processes`` sets the number of processes to use. This should
   typically be set to the number of CPU threads, but gains in speed are minimal

options.cpp

@@ -227,7 +227,8 @@ namespace
             printf("    %s -c -d gis --slim -C <cache size> -k \\\n", name);
             printf("      --flat-nodes <flat nodes> planet-latest.osm.pbf\n");
             printf("where\n");
-            printf("    <cache size> is 50000 on machines with 64GB or more RAM \n");
+            printf("    <cache size> should be equivalent to the size of the \n");
+            printf("      pbf file to be imported if there is enough RAM \n");
             printf("      or about 75%% of memory in MB on machines with less\n");
             printf("    <flat nodes> is a location where a 50+GB file can be saved.\n");
             printf("\n");