We had to restart a server after a reboot of the Virtual machine the Rabbitmq server is working on.
The server was properly shutdown :
and then at the restart 
=INFO REPORT==== 29-Apr-2015::13:56:56 ===
Starting RabbitMQ 3.4.2 on Erlang 17.1
Copyright (C) 2007-2014 GoPivotal, Inc.
Licensed under the MPL.  See 
http://www.rabbitmq.com/=INFO REPORT==== 29-Apr-2015::13:56:56 ===
node           : nodename@machinename
home dir       : /home/rabbitmq
config file(s) : /PATHTOCONFIG/rabbitmq/rabbitmq.config
cookie hash    : UHdCqMlg8zjiF0qecJzwrQ==
log            : /PATHTOTHELOG/LOGNAME.log
sasl log       :  /PATHTOTHELOG/LOGNAME-sasl.log
database dir   : /PATHTODATABASE
=INFO REPORT==== 29-Apr-2015::13:56:57 ===
Limiting to approx 64900 file handles (58408 sockets)
=INFO REPORT==== 29-Apr-2015::13:58:52 ===
Memory limit set to 13994MB of 19991MB total.
=INFO REPORT==== 29-Apr-2015::13:58:52 ===
Disk free limit set to 50MB
=INFO REPORT==== 29-Apr-2015::13:58:52 ===
msg_store_transient: using rabbit_msg_store_ets_index to provide index
=INFO REPORT==== 29-Apr-2015::13:58:52 ===
msg_store_persistent: using rabbit_msg_store_ets_index to provide index
=WARNING REPORT==== 29-Apr-2015::13:58:52 ===
msg_store_persistent: rebuilding indices from scratch
This step tooks 1 hour and 30 minutes to complete before to complete the start :
=INFO REPORT==== 29-Apr-2015::15:26:13 ===
started TCP Listener on 
0.0.0.0:51012=INFO REPORT==== 29-Apr-2015::15:26:14 ===
Management plugin started. Port: 52012
=INFO REPORT==== 29-Apr-2015::15:26:14 ===
Server startup complete; 8 plugins started.
 * rabbitmq_management
 * rabbitmq_web_dispatch
 * webmachine
 * mochiweb
 * rabbitmq_auth_backend_ldap
 * rabbitmq_management_agent
 * amqp_client
 * eldap
 Could it be possible to avoid such  "rebuilding indices from scratch" how to make sure this is not done each time we have to stop and start the server ? 
Many thanks in advance