The NetApp Docker Volume Plugin (nDVP) provides direct integration with the Docker ecosystem for NetApp's ONTAP, E-Series, and SolidFire storage platforms. The nDVP package supports the provisioning and management of storage resources from the storage platform to Docker hosts, with a robust framework for adding additional platforms in the future.

Multiple instances of the nDVP can run concurrently on the same host. The allows simultaneous connections to multiple storage systems and storage types, with the ablity to customize the storage used for the Docker volume(s).

• Quick Start
• Running Multiple nDVP Instances
• Configuring your Docker host for NFS or iSCSI
  • NFS
    • RHEL / CentOS
    • Ubuntu / Debian
  • iSCSI
    • RHEL / CentOS
    • Ubuntu / Debian
• Global Config File Variables
  • Storage Prefix
• ONTAP Config File Variables
  • Example ONTAP Config Files
• E-Series Config File Variables
  • Example E-Series Config Files
  • E-Series Array Setup Notes
• SolidFire Config File Variables
  • Example Solidfire Config Files

1. Ensure you have Docker version 1.10 or above.

   docker --version

   If your version is out of date, update to the latest.

   curl -fsSL https://get.docker.com/ | sh

   Or, follow the instructions for your distribution.

2. After ensuring the correct version of Docker is installed, install and configure the NetApp Docker Volume Plugin. Note, you will need to ensure that NFS and/or iSCSI is configured for your system. See the installation instructions below for detailed information on how to do this.

wget -O /tmp/netappdvp-1.4.0.tar.gz https://github.com/NetApp/netappdvp/releases/download/v1.4.0/netappdvp-1.4.0.tar.gz cd /tmp tar zxf netappdvp-1.4.0.tar.gz

sudo mv netappdvp /usr/local/bin sudo chown root:root /usr/local/bin/netappdvp sudo chmod 755 /usr/local/bin/netappdvp

sudo mkdir -p /etc/netappdvp

cat << EOF > /etc/netappdvp/ontap-nas.json { "version": 1, "storageDriverName": "ontap-nas", "managementLIF": "10.0.0.1", "dataLIF": "10.0.0.2", "svm": "docker_svm", "username": "vsadmin", "password": "password", "aggregate": "aggr", "storagePrefix": "netappdvp_", } EOF ```

3. After placing the binary and creating the configuration file(s), start the nDVP daemon using the desired configuration file.

   Note: Unless specified, the default name for the volume driver will be "netapp".

   sudo netappdvp --config=/etc/netappdvp/ontap-nas.json

4. Once the daemon is started, create and manage volumes using the Docker CLI interface.

   docker volume create -d netapp --name ndvp_1

   Provision Docker volume when starting a container:

   docker run --rm -it --volume-driver netapp --volume ndvp_2:/my_vol alpine ash

   Destroy docker volume:

   docker volume rm ndvp_1
   docker volume rm ndvp_2

1. Launch the plugin with an NFS configuration using a custom driver ID:

   sudo netappdvp --volume-driver=netapp-nas --config=/path/to/config-nfs.json

2. Launch the plugin with an iSCSI configuration using a custom driver ID:

   sudo netappdvp --volume-driver=netapp-san --config=/path/to/config-iscsi.json

3. Provision Docker volumes each driver instance:

   NFS

   docker volume create -d netapp-nas --name my_nfs_vol

   iSCSI

   docker volume create -d netapp-san --name my_iscsi_vol

1. Create the netappdvp systemd service

   systemd

   cat << EOF > /etc/systemd/system/netappdvp.service
   [Unit]
   Description=Start Netapp Docker Volume Plugin
   Before=docker.service
   After=network.target
   Requires=docker.service
   
   [Service]
   ExecStart=/etc/netappdvp/netappdvp_startup.sh
   
   [Install]
   WantedBy=multi-user.target
   EOF

   configure startup options

   sudo systemctl enable /etc/systemd/system/netappdvp.service
   sudo systemctl daemon-reload

2. Create the startup script

   bash script

   cat << EOF > /etc/netappdvp/netappdvp_startup.sh
   #!/bin/bash
   /usr/local/bin/netappdvp --config=/etc/netappdvp/ontap-nas.json
   EOF
   
   chmod +x /etc/netappdvp/netappdvp_startup.sh

restart services bash systemctl restart netappdvp systemctl restart docker

Install the following system packages:

sudo yum install -y nfs-utils

sudo apt-get install -y nfs-common

Other things needed

• Add in an iSCSI NIC (this only handles a single source NIC, so make it a failover NIC)
• Configure an IP address on your iSCSI NIC
• get the name of your iSCSI NIC 'connection' with nmcli con show

Fill out the variables below and the rest is pretty much blind copy and paste.

#some variables to smooth out the process, change the areas contained in <foo>  
ISCSI_NIC_NAME="<interfacename>"  
SVM_NAME="<SVM name>"  
SVM_MGMT_IP="<SVM management IP>"  
SVM_ISCSI_IP1="<SVM iSCSI LIF IP1>"  
SVM_ISCSI_IP2="<SVM iSCSI LIF IP2>"  
VSADMIN_PASSWORD="<password>"  
AGGREGATE_NAME="<aggregate name>"  
STORAGE_PREFIX="<storage prefix>"
IGROUP="$(hostname -s)"  
  
#install dependencies  
sudo yum install -y lsscsi iscsi-initiator-utils sg3_utils device-mapper-multipath

#append hostname to IQN so its easy to identify
sudo printf "$(< /etc/iscsi/initiatorname.iscsi)-$(hostname -s)\n" > /etc/iscsi/initiatorname.iscsi

#configure iscsi virtual interface and bind it to the iscsi nics mac address
sudo iscsiadm -m iface -I ieth1 --op=new
sudo iscsiadm -m iface -I ieth1 --op=update -n iface.net_ifacename -v ${ISCSI_NIC_NAME}

#change iscsi interface to have a 9k mtu and not submit dns registration
sudo nmcli con mod ${ISCSI_NIC_CON_NAME} 802-3-ethernet.mtu 9000
sudo nmcli device mod ${ISCSI_NIC_NAME} ipv6.dhcp-send-hostname no
sudo nmcli device mod ${ISCSI_NIC_NAME} ipv4.dhcp-send-hostname no

#enable multipathd
sudo mpathconf --enable --with_multipathd y
sudo systemctl enable iscsid multipathd
sudo systemctl start iscsid multipathd

#change iscsi session timeout per netapps recommended best practice 
sudo sed -i 's/node.session.timeo.replacement_timeout = 120/node.session.timeo.replacement_timeout = 5/' /etc/iscsi/iscsid.conf

#configure discovery for both iscsi IPs
sudo iscsiadm -m discoverydb -t st -p ${SVM_ISCSI_IP1} --discover -I ieth1
sudo iscsiadm -m discoverydb -t st -p ${SVM_ISCSI_IP2} --discover -I ieth1
sudo iscsiadm -m node -p ${SVM_ISCSI_IP1} -I ieth1 --login
sudo iscsiadm -m node -p ${SVM_ISCSI_IP2} -I ieth1 --login

#enable iscsi
sudo systemctl enable iscsi
sudo systemctl start iscsi

1. Install the following system packages:

   sudo apt-get install -y open-iscsi lsscsi sg3-utils multipath-tools scsitools

2. Enable multipathing:

   sudo tee /etc/multipath.conf <<-'EOF'
   defaults {
       user_friendly_names yes
       find_multipaths yes
   }
   EOF
   
   sudo service multipath-tools restart

3. Ensure that iscsid and multipathd are running:

   sudo service open-iscsi start
   sudo service multipath-tools start

4. Discover the iSCSI targets:

   sudo iscsiadm -m discoverydb -t st -p <DATA_LIF_IP> --discover

5. Login to the discovered iSCSI targets:

   sudo iscsiadm -m node -p <DATA_LIF_IP> --login

A new config file variable has been added in v1.2 called "storagePrefix" that allows you to modify the prefix applied to volume names by the plugin. By default, when you run docker volume create, the volume name supplied is prepended with "netappdvp_". ("netappdvp-" for SolidFire.)

If you wish to use a different prefix, you can specify it with this directive. Alternatively, you can use pre-existing volumes with the volume plugin by setting storagePrefix to an empty string, "".

Ontap specific recommendation Keeping the same storagePrefix across multiple docker hosts will show all docker volumes created with the same storagePrefix, even if you use a different igroup. Use a different storagePrefix per docker host barring shared volumes. With iSCSI you are unlikely to ever want this scenario to happen.

solidfire specific recommendation do not use a storagePrefix (including the default) By default the SolidFire driver will ignore this setting and not use a prefix. We recommend using either a specific tenantID for docker volume mapping or using the attribute data which is populated with the docker version, driver info and raw name from docker in cases where any name munging may have been used.

A note of caution: If you can see a volume with docker volume ls due to your storagePrefix name, then docker volume rm <volumename> will delete the specified volume, even if you aren't in the same igroup or if it was not created by the netappdvp plugin! Be very careful when using pre-existing volumes!

In addition to the global configuration values above, when using clustered Data ONTAP, these options are avaialble.

NFS Example for ontap-nas driver

{
    "version": 1,
    "storageDriverName": "ontap-nas",
    "managementLIF": "10.0.0.1",
    "dataLIF": "10.0.0.2",
    "svm": "svm_nfs",
    "username": "vsadmin",
    "password": "netapp123",
    "aggregate": "aggr1",
    "storagePrefix": "mydocker_nfs_"
}

iSCSI Example for ontap-san driver

{
    "version": 1,
    "storageDriverName": "ontap-san",
    "managementLIF": "10.0.0.1",
    "dataLIF": "10.0.0.3",
    "svm": "svm_iscsi",
    "username": "vsadmin",
    "password": "netapp123",
    "aggregate": "aggr1",
    "IgroupName": "mydockerhost",
    "storagePrefix": "mydocker_iscsi_"
}

In addition to the global configuration values above, when using E-Series, these options are available.

Example for eseries-iscsi driver

{
	"version": 1,
	"storageDriverName": "eseries-iscsi",
	"debug": true,
	"webProxyHostname": "localhost",
	"webProxyPort": "8443",
	"webProxyUseHTTP": false,
	"webProxyVerifyTLS": true,
	"username": "rw",
	"password": "rw",
	"controllerA": "10.0.0.5",
	"controllerB": "10.0.0.6",
	"passwordArray": "",
	"hostDataIP": "10.0.0.101"
}

The E-Series Docker driver can provision Docker volumes in any storage pool on the array, including volume groups and DDP pools. To limit the Docker driver to a subset of the storage pools, set the poolNameSearchPattern in the configuration file to a regular expression that matches the desired pools.

When creating a docker volume you can specify the pool name, volume size, and disk media type using the '-o' option and the tags 'pool, 'size', and 'mediaType', respectively. Valid values for media type are 'hdd' and 'ssd'. Note that these are optional; if unspecified, the defaults will be a 1 GB volume allocated from an HDD pool. An example of using these tags to create a 2 GB volume from any available SSD-based pool with sufficient space available:

docker volume create -d netapp --name my_vol -o size=2g -o mediaType=ssd

The E-series Docker driver will detect and use any preexisting Host definitions without modification, and the driver will automatically define Host and Host Group objects as needed. The host type for hosts created by the driver defaults to "linux_dm_mp", the native DM-MPIO multipath driver in Linux.

The current E-series Docker driver only supports iSCSI.

In addition to the global configuration values above, when using SolidFire, these options are available.

LegacyNamePrefix If you used a version of ndvp prior to 1.3.2 and perform an upgrade with existing volumes, you'll need to set this value in order to access your old volumes that were mapped via the volume-name method.

{
    "version": 1,
    "storageDriverName": "solidfire-san",
    "debug": false,
    "Endpoint": "https://admin:[email protected]/json-rpc/7.0",
    "SVIP": "10.0.0.7:3260",
    "TenantName": "docker",
    "DefaultVolSz": 1,
    "InitiatorIFace": "default",
    "Types": [
        {
            "Type": "Bronze",
            "Qos": {
                "minIOPS": 1000,
                "maxIOPS": 2000,
                "burstIOPS": 4000
            }
        },
        {
            "Type": "Silver",
            "Qos": {
                "minIOPS": 4000,
                "maxIOPS": 6000,
                "burstIOPS": 8000
            }
        },
        {
            "Type": "Gold",
            "Qos": {
                "minIOPS": 6000,
                "maxIOPS": 8000,
                "burstIOPS": 10000
            }
        }
    ]

}

##Known Issues and Limitations

1. Volume names must be a minimum of 2 characters in length

   This is a Docker client limitation. The client will interpret a single character name as being a Windows path. Bug 25773