Upraviť

Zdieľať cez

Configure an NFS client for Azure NetApp Files

  • Článok

The NFS client configuration described in this article is part of the setup when you configure NFSv4.1 Kerberos encryption or create a dual-protocol volume or NFSv3/NFSv4.1 with LDAP. A wide variety of Linux distributions are available to use with Azure NetApp Files. This article describes configurations for two of the more commonly used environments: RHEL 8 and Ubuntu 18.04.

Requirements and considerations

Regardless of the Linux flavor you use, the following configurations are required:

  • Configure an NTP client to avoid issues with time skew.
  • Configure DNS entries of the Linux client for name resolution.
    This configuration must include the “A” (forward) record and the PTR (reverse) record .
  • For domain join, create a computer account for the Linux client in the target Active Directory (which is created during the realm join command).

    Note

    The $SERVICEACCOUNT variable used in the commands below should be a user account with permissions or delegation to create a computer account in the targeted Organizational Unit.

RHEL 8 configuration

This section describes RHEL configurations required for NFSv4.1 Kerberos encryption and dual protocol.

The examples in this section use the following domain name and IP address:

  • Domain name: contoso.com
  • Private IP: 10.6.1.4

RHEL 8 configuration if you are using NFSv4.1 Kerberos encryption

  1. Configure /etc/resolv.conf with the proper DNS server.

    For example:

    [root@reddoc cbs]# cat /etc/resolv.conf
    search contoso.com
    nameserver 10.6.1.4(private IP)

  2. Add the NFS client record in the DNS server for the DNS forward and reverse lookup zone.

  3. To verify DNS, use the following commands from the NFS client:

    # nslookup [hostname/FQDN of NFS client(s)]
    # nslookup [IP address of NFS client(s)]

  4. Install packages:

    yum update
    sudo yum -y install realmd sssd adcli samba-common krb5-workstation chrony nfs-utils

  5. Configure the NTP client.

    RHEL 8 uses chrony by default.

  6. Join the Active Directory domain:

    sudo realm join $DOMAIN.NAME -U $SERVICEACCOUNT --computer-ou="OU=$YOUROU"

    For example:

    sudo realm join CONTOSO.COM -U ad_admin --computer-ou="CN=Computers"

    Ensure that default_realm is set to the provided realm in /etc/krb5.conf. If not, add it under the [libdefaults] section in the file as shown in the following example:

    [libdefaults]
    default_realm = CONTOSO.COM
    default_tkt_enctypes = aes256-cts-hmac-sha1-96
    default_tgs_enctypes = aes256-cts-hmac-sha1-96
    permitted_enctypes = aes256-cts-hmac-sha1-96
[realms]
    CONTOSO.COM = {
        kdc = dc01.contoso.com
        admin_server = dc01.contoso.com
        master_kdc = dc01.contoso.com
        default_domain = contoso.com
    }
[domain_realm]
    .contoso.com = CONTOSO.COM
    contoso.com = CONTOSO.COM
[logging]
    kdc = SYSLOG:INFO
    admin_server = FILE=/var/kadm5.log

  7. Restart all NFS services:

    systemctl start nfs-*
    systemctl restart rpc-gssd.service

    The restart prevents the error condition “mount.nfs: an incorrect mount option was specified” during Kerberos mount.

  8. Run the kinit command with the user account to get tickets:

    sudo kinit $SERVICEACCOUNT@DOMAIN

    For example:

    sudo kinit ad_admin@CONTOSO.COM

RHEL 8 configuration if you are using dual protocol

The following steps are optional. You need to perform the steps only if you use user mapping at the NFS client:

  1. Complete all steps described in the RHEL 8 configuration if you are using NFSv4.1 Kerberos encryption section.

  2. Add a static DNS record in your /etc/hosts file to use fully qualified domain name (FQDN) for your AD, instead of using the IP address in SSSD configuration file:

    cat /etc/hosts
    10.6.1.4 winad2016.contoso.com

  3. Add an extra section for domains to resolve identifiers from AD LDAP server:

    [root@reddoc cbs]# cat /etc/sssd/sssd.conf
    [sssd]
    domains = contoso.com, contoso-ldap (new entry added for LDAP as id_provider)
    config_file_version = 2
    services = nss, pam, ssh, sudo (ensure nss is present in this list)

    [domain/contoso-ldap] (Copy the following lines. Modify as per your domain name.)
    auth_provider = krb5
    chpass_provider = krb5
    id_provider = ldap
    ldap_search_base = dc=contoso,dc=com(your domain)
    ldap_schema = rfc2307bis
    ldap_sasl_mech = GSSAPI
    ldap_user_object_class = user
    ldap_group_object_class = group
    ldap_user_home_directory = unixHomeDirectory
    ldap_user_principal = userPrincipalName
    ldap_account_expire_policy = ad
    ldap_force_upper_case_realm = true
    ldap_user_search_base = cn=Users,dc=contoso,dc=com (based on your domain)
    ldap_group_search_base = cn=Users,dc=contoso,dc=com (based on your domain)
    ldap_sasl_authid = REDDOC$ (ensure $ at the end you can get this from “klist -kte” command)
    krb5_server = winad2016.contoso.com (same as AD address which is added in /etc/hosts)
    krb5_realm = CONTOSO.COM (domain name in caps)
    krb5_kpasswd = winad2016.contoso.com (same as AD address which is added in /etc/hosts)
    use_fully_qualified_names = false

    In the [domain/contoso-ldap] configuration above:

    • id_provider is set to ldap and not ad.
    • The configuration has specified search bases and user and group classes for searches.
    • ldap_sasl_authid is the machine account name from klist -kte.
    • use_fully_qualified_names is set to false. This setting means this configuration is used when a short name is used.
    • ldap_id_mapping is NOT specified, which defaults to false.

    The realm join configuration is generated by the client and looks like this:

    [domain/contoso.com] (Do not edit or remove any of the following information. This information is automatically generated during the realm join process.)
    ad_domain = contoso.com
    krb5_realm = CONTOSO.COM
    realmd_tags = manages-system joined-with-adcli
    cache_credentials = True
    id_provider = ad
    krb5_store_password_if_offline = True
    default_shell = /bin/bash
    ldap_id_mapping = True
    use_fully_qualified_names = True
    fallback_homedir = /home/%u@%d
    access_provider = ad

    In the [domain/contoso.com] configuration above:

    • id_provider is set to ad.
    • ldap_id_mapping is set to true. It uses the SSSD generated IDs. Alternately, you can set this value to false if you want to use POSIX UIDs for ALL styles of usernames. You can determine the value based on your client configuration.
    • use_fully_qualified_names is true. This setting means user@CONTOSO.COM will use this configuration.

  4. Ensure your /etc/nsswitch.conf has the sss entry:

    cat /etc/nsswitch.conf
    passwd: sss files systemd
    group: sss files systemd
    netgroup: sss files

  5. Restart the sssd service and clear cache:

    service sssd stop
    rm -f /var/lib/sss/db/*
    service sssd start

  6. Test to ensure that your client is integrated with the LDAP server:

    [root@red81 cbs]# id ldapuser1
    uid=1234(ldapuser1) gid=1111(ldapgroup1) groups=1111(ldapgroup1)

Ubuntu configuration

This section describes Ubuntu configurations required for NFSv4.1 Kerberos encryption and dual protocol.

The examples in this section use the following domain name and IP address:

  • Domain name: contoso.com
  • Private IP: 10.6.1.4

  1. Configure /etc/resolv.conf with the proper DNS server:

    root@ubuntu-rak:/home/cbs# cat /etc/resolv.conf
    search contoso.com
    nameserver <private IP address of DNS server>

  2. Add NFS client record in the DNS server for the DNS forward and reverse lookup zone.

    To verify DNS, use the following commands from the NFS client:

    # nslookup [hostname/FQDN of NFS client(s)]
    # nslookup [IP address of NFS client(s)]

  3. Install packages:

    apt-get update
    apt-get install -y realmd packagekit sssd adcli samba-common chrony krb5-user nfs-common

    When prompted, input $DOMAIN.NAME (using uppercase, for example, CONTOSO.COM) as the default Kerberos realm.

  4. Restart the service rpc-gssd.service:

    sudo systemctl start rpc-gssd.service

  5. Ubuntu 18.04 uses chrony by default. Following the configuration guidelines in Ubuntu Bionic: Using chrony to configure NTP.

  6. Join the Active Directory domain:

    sudo realm join $DOMAIN.NAME -U $SERVICEACCOUNT --computer-ou="OU=$YOUROU"

    For example:
    sudo realm join CONTOSO.COM -U ad_admin --computer-ou="CN=Computers"

  7. Perform kinit with the user to get tickets:

    sudo kinit $SERVICEACCOUNT

    For example:
    sudo kinit ad_admin

Ubuntu configuration if you are using dual protocol

The following steps are optional. You need to perform the steps only if you want to use user mapping at the NFS client:

  1. Run the following command to upgrade the installed packages:
    sudo apt update && sudo apt install libnss-ldap libpam-ldap ldap-utils nscd

    The following example uses sample values. When the command prompts you for input, you should provide input based on your environment.

    base dc=contoso,dc=com uri ldap://10.20.0.4:389/ ldap_version 3 rootbinddn cn=admin,cn=Users,dc=contoso,dc=com pam_password ad

  2. Ensure that your /etc/nsswitch.conf file has the following ldap entries:
    passwd: compat systemd ldap
    group: compat systemd ldap

  3. Run the following command to restart and enable the service:

    sudo systemctl restart nscd && sudo systemctl enable nscd

The following example queries the AD LDAP server from Ubuntu LDAP client for an LDAP user ‘hari1’:

root@cbs-k8s-varun4-04:/home/cbs# getent passwd hari1
hari1:*:1237:1237:hari1:/home/hari1:/bin/bash

Configure two VMs with the same hostname to access NFSv4.1 volumes

This section explains how you can configure two VMs that have the same hostname to access Azure NetApp Files NFSv4.1 volumes. This procedure can be useful when you conduct a disaster recovery (DR) test and require a test system with the same hostname as the primary DR system. This procedure is only required when you have the same hostname on two VMs that are accessing the same Azure NetApp Files volumes.

NFSv4.x requires each client to identify itself to servers with a unique string. File open and lock state shared between one client and one server is associated with this identity. To support robust NFSv4.x state recovery and transparent state migration, this identity string must not change across client reboots.

  1. Display the nfs4_unique_id string on the VM clients by using the following command:

    # systool -v -m nfs | grep -i nfs4_unique
    nfs4_unique_id = ""

    To mount the same volume on an additional VM with the same hostname, for example the DR system, create a nfs4_unique_id so it can uniquely identify itself to the Azure NetApp Files NFS service. This step allows the service to distinguish between the two VMs with the same hostname and enable mounting NFSv4.1 volumes on both VMs.

    You need to perform this step on the test DR system only. For consistency, you can consider applying a unique setting on each involved virtual machine.

  2. On the test DR system, add the following line to the nfsclient.conf file, typically located in /etc/modprobe.d/:

    options nfs nfs4_unique_id=uniquenfs4-1

    The string uniquenfs4-1 can be any alphanumeric string, as long as it is unique across the VMs to be connected to the service.

    Check your distribution’s documentation about how to configure NFS client settings.

    Reboot the VM for the change to take effect.

  3. On the test DR system, verify that nfs4_unique_id has been set after the VM reboot:

    # systool -v -m nfs | grep -i nfs4_unique
    nfs4_unique_id = "uniquenfs4-1"

  4. Mount the NFSv4.1 volume on both VMs as normal.

    Both VMs with the same hostname can now mount and access the NFSv4.1 volume.

Next steps