diff --git a/README.md b/README.md
index 480d71e..4ed18e7 100644
--- a/README.md
+++ b/README.md
@@ -60,14 +60,18 @@ You can build each parts separatly by running `make` in each separated directori
 # Running
 
 0-db is made to be run in network server mode (using zdbd), documentation here is about the server.
-More documentation will comes about the library itself. The library is fresh new and lack of documentation.
-
-## Default port
-0-db listens by default on port `9900` but this can be overidden on the commandline using `--port` option.
+More documentation will comes about the library itself. The library lacks of documentation.
 
 Without argument, datafiles and indexfiles will be stored on the current working directory, inside
 `zdb-data` and `zdb-index` directories.
 
+It's recommended to store index on SSD, data can be stored on HDD, fragmentation will be avoided as much
+as possible (only on Linux). Please avoid CoW (Copy-on-Write) for both index and data.
+
+## Default port
+
+0-db listens by default on port `9900` but this can be overidden on the commandline using `--port` option.
+
 # Always append
 Data file (files which contains everything, included payload) are **in any cases** always append:
 any change will result in something appened to files. Data files are immuables. If any suppression is
diff --git a/libzdb/data.c b/libzdb/data.c
index 54c8d47..247f700 100644
--- a/libzdb/data.c
+++ b/libzdb/data.c
@@ -1,3 +1,5 @@
+#define _GNU_SOURCE
+
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -251,6 +253,13 @@ void data_initialize(char *filename, data_root_t *root) {
         zdb_diep(filename);
     }
 
+    #ifdef __linux__
+    // pre-allocate data size to avoid fragmentation, only on Linux
+    zdb_settings_t *settings = zdb_settings_get();
+    if(fallocate(fd, FALLOC_FL_KEEP_SIZE, 0, settings->datasize) < 0)
+        zdb_warnp(filename);
+    #endif
+
     // writing initial header
     data_header_t header;