diff --git a/doc/guide.html b/doc/guide.html
index b2144167d..0a21eef5c 100644
--- a/doc/guide.html
+++ b/doc/guide.html
@@ -3417,11 +3417,13 @@ they are automatically renamed to <TT>"*-old.log"</TT>. See section <A HREF="#lo
 Store internal Mnesia database to a binary backup file.
 </DD><DT CLASS="dt-description"><B><TT>restore ejabberd.backup</TT></B></DT><DD CLASS="dd-description">
 Restore immediately from a binary backup file the internal Mnesia database.
-This will consume quite some memory for big servers.
+This will consume a lot of memory if you have a large database,
+so better use <TT>install_fallback</TT>.
 </DD><DT CLASS="dt-description"><B><TT>install_fallback ejabberd.backup</TT></B></DT><DD CLASS="dd-description">
 The binary backup file is installed as fallback:
 it will be used to restore the database at the next ejabberd start.
-Similar to <TT>restore</TT>, but requires less memory.
+This means that, after running this command, you have to restart ejabberd.
+This command requires less memory than <TT>restore</TT>.
 </DD><DT CLASS="dt-description"><B><TT>dump ejabberd.dump</TT></B></DT><DD CLASS="dd-description">
 Dump internal Mnesia database to a text file dump.
 </DD><DT CLASS="dt-description"><B><TT>load ejabberd.dump</TT></B></DT><DD CLASS="dd-description">
@@ -3604,8 +3606,17 @@ mv /var/lib/ejabberd/*.* /var/lib/ejabberd/oldfiles/
 <PRE CLASS="verbatim">ejabberdctl start
 </PRE></LI><LI CLASS="li-enumerate">Convert the backup to new node name:
 <PRE CLASS="verbatim">ejabberdctl mnesia_change_nodename $OLDNODE $NEWNODE $OLDFILE $NEWFILE
-</PRE></LI><LI CLASS="li-enumerate">Import the new backup:
-<PRE CLASS="verbatim">ejabberdctl restore $NEWFILE
+</PRE></LI><LI CLASS="li-enumerate">Install the backup file as a fallback:
+<PRE CLASS="verbatim">ejabberdctl install_fallback $NEWFILE
+</PRE></LI><LI CLASS="li-enumerate">Stop ejabberd:
+<PRE CLASS="verbatim">ejabberdctl stop
+</PRE>You may see an error message in the log files, it&#X2019;s normal, so don&#X2019;t worry:
+<PRE CLASS="verbatim">Mnesia(ejabberd@newmachine):
+** ERROR ** (ignoring core)
+** FATAL ** A fallback is installed and Mnesia must be restarted.
+  Forcing shutdown after mnesia_down from ejabberd@newmachine...
+</PRE></LI><LI CLASS="li-enumerate">Now you can finally start ejabberd:
+<PRE CLASS="verbatim">ejabberdctl start
 </PRE></LI><LI CLASS="li-enumerate">Check that the information of the old database is available: accounts, rosters...
 After you finish, remember to delete the temporary backup files from public directories.
 </LI></OL><P> <A NAME="secure"></A> </P><!--TOC chapter Securing <TT>ejabberd</TT>-->
diff --git a/doc/guide.tex b/doc/guide.tex
index 71d830bbf..edb626c43 100644
--- a/doc/guide.tex
+++ b/doc/guide.tex
@@ -4364,11 +4364,13 @@ The most interesting ones are:
   Store internal Mnesia database to a binary backup file.
 \titem {restore ejabberd.backup}
   Restore immediately from a binary backup file the internal Mnesia database.
-  This will consume quite some memory for big servers.
+  This will consume a lot of memory if you have a large database,
+  so better use \term{install\_fallback}.
 \titem {install\_fallback ejabberd.backup}
   The binary backup file is installed as fallback:
   it will be used to restore the database at the next ejabberd start.
-  Similar to \term{restore}, but requires less memory.
+  This means that, after running this command, you have to restart ejabberd.
+  This command requires less memory than \term{restore}.
 \titem {dump ejabberd.dump}
   Dump internal Mnesia database to a text file dump.
 \titem {load ejabberd.dump}
@@ -4617,9 +4619,26 @@ ejabberdctl start
 ejabberdctl mnesia_change_nodename $OLDNODE $NEWNODE $OLDFILE $NEWFILE
 \end{verbatim}
 
-\item Import the new backup:
+\item Install the backup file as a fallback:
 \begin{verbatim}
-ejabberdctl restore $NEWFILE
+ejabberdctl install_fallback $NEWFILE
+\end{verbatim}
+
+\item Stop ejabberd:
+\begin{verbatim}
+ejabberdctl stop
+\end{verbatim}
+You may see an error message in the log files, it's normal, so don't worry:
+\begin{verbatim}
+Mnesia(ejabberd@newmachine):
+** ERROR ** (ignoring core)
+** FATAL ** A fallback is installed and Mnesia must be restarted.
+  Forcing shutdown after mnesia_down from ejabberd@newmachine...
+\end{verbatim}
+
+\item Now you can finally start ejabberd:
+\begin{verbatim}
+ejabberdctl start
 \end{verbatim}
 
 \item Check that the information of the old database is available: accounts, rosters...