Configure an NFS storage back end¶
This section explains how to configure OpenStack Block Storage to use
NFS storage. You must be able to access the NFS shares from the server
that hosts the cinder volume service.
Note
The cinder volume service is named openstack-cinder-volume
on the following distributions:
- CentOS 
- Fedora 
- openSUSE 
- Red Hat Enterprise Linux 
- SUSE Linux Enterprise 
In Ubuntu and Debian distributions, the cinder volume service is
named cinder-volume.
Configure Block Storage to use an NFS storage back end
- Log in as - rootto the system hosting the- cindervolume service.
- Create a text file named - nfs_sharesin the- /etc/cinder/directory.
- Add an entry to - /etc/cinder/nfs_sharesfor each NFS share that the- cindervolume service should use for back end storage. Each entry should be a separate line, and should use the following format:- HOST:SHARE - Where: - HOST is the IP address or host name of the NFS server. 
- SHARE is the absolute path to an existing and accessible NFS share. 
 
- Set - /etc/cinder/nfs_sharesto be owned by the- rootuser and the- cindergroup:- # chown root:cinder /etc/cinder/nfs_shares 
- Set - /etc/cinder/nfs_sharesto be readable by members of the cinder group:- # chmod 0640 /etc/cinder/nfs_shares 
- Configure the - cindervolume service to use the- /etc/cinder/nfs_sharesfile created earlier. To do so, open the- /etc/cinder/cinder.confconfiguration file and set the- nfs_shares_configconfiguration key to- /etc/cinder/nfs_shares.- On distributions that include - openstack-config, you can configure this by running the following command instead:- # openstack-config --set /etc/cinder/cinder.conf \ DEFAULT nfs_shares_config /etc/cinder/nfs_shares - The following distributions include openstack-config: - CentOS 
- Fedora 
- openSUSE 
- Red Hat Enterprise Linux 
- SUSE Linux Enterprise 
 
- Optionally, provide any additional NFS mount options required in your environment in the - nfs_mount_optionsconfiguration key of- /etc/cinder/cinder.conf. If your NFS shares do not require any additional mount options (or if you are unsure), skip this step.- On distributions that include - openstack-config, you can configure this by running the following command instead:- # openstack-config --set /etc/cinder/cinder.conf \ DEFAULT nfs_mount_options OPTIONS - Replace OPTIONS with the mount options to be used when accessing NFS shares. See the manual page for NFS for more information on available mount options (man nfs). 
- Configure the - cindervolume service to use the correct volume driver, namely- cinder.volume.drivers.nfs.NfsDriver. To do so, open the- /etc/cinder/cinder.confconfiguration file and set the volume_driver configuration key to- cinder.volume.drivers.nfs.NfsDriver.- On distributions that include - openstack-config, you can configure this by running the following command instead:- # openstack-config --set /etc/cinder/cinder.conf \ DEFAULT volume_driver cinder.volume.drivers.nfs.NfsDriver 
- You can now restart the service to apply the configuration. - Note - The - nfs_sparsed_volumesconfiguration key determines whether volumes are created as sparse files and grown as needed or fully allocated up front. The default and recommended value is- true, which ensures volumes are initially created as sparse files.- Setting - nfs_sparsed_volumesto- falsewill result in volumes being fully allocated at the time of creation. This leads to increased delays in volume creation.- However, should you choose to set - nfs_sparsed_volumesto- false, you can do so directly in- /etc/cinder/cinder.conf.- On distributions that include - openstack-config, you can configure this by running the following command instead:- # openstack-config --set /etc/cinder/cinder.conf \ DEFAULT nfs_sparsed_volumes false - Warning - If a client host has SELinux enabled, the - virt_use_nfsboolean should also be enabled if the host requires access to NFS volumes on an instance. To enable this boolean, run the following command as the- rootuser:- # setsebool -P virt_use_nfs on - This command also makes the boolean persistent across reboots. Run this command on all client hosts that require access to NFS volumes on an instance. This includes all compute nodes. 
