Configure open NFS ports in AWS EKS
SharedV4 volumes utilize NFS services, and they therefore require specific open NFS ports to allow for communication between nodes in your cluster. Depending on how your cluster nodes are configured, your firewall may block some of these ports, or your NFS ports may differ from the defaults. To solve these issues, you may need to manually assign NFS ports and ensure that your firewall or ACL allows them to communicate.
Portworx supports the following types of ReadWriteMany volumes:
If a PersistentVolumeClaim (PVC) has ReadWriteMany (RWX) access mode, Portworx creates a sharedv4 service volume by default. Sharedv4 service volumes use NFSv4 by default, which means only port 2049 needs to be opened. However, sharedv4 volumes use NFSv3 by default. If your volumes use another port, see the Manually assign and open NFS ports section.
This document provides instructions for detecting and opening NFS ports according to various cluster configurations that you may have.
Prerequisites
All of the use-cases in this document will require that the mandatory Portworx network port ranges are open between nodes in the cluster, as documented here.
Determine which ports to open
First, check what the existing NFS port configuration is for your nodes to see if they need to be remapped.
Enter the following command to find which ports NFS is using on your node:
rpcinfo -p
SharedV4 volumes communicate on the following standard ports/services:
- PortMapper: tcp/udp 111 (default on most Linux distributions)
- NFSd: tcp/udp 2049 (default on most Linux distributions)
- MountD: tcp/udp 20048 (depends on the Linux distribution)
If the ports listed from the above rpcinfo
output on your nodes match these standard ports, proceed to Open standard NFS ports (most Linux distributions).
If the NFS ports on your OS do not match these ports, or your OS randomly chooses the ports for these services, proceed to Manually assign and open NFS ports.
Open standard NFS ports (most Linux distributions)
If your Linux distribution uses the standard ports identified in the previous section, you do not need to manually assign any ports for NFS, but you may need to open them.
Ensure that your ports are open on any firewalls and your ACL by entering the following commands:
iptables -I INPUT -p tcp -m tcp --match multiport --dports 111,2049,20048 -j ACCEPT
iptables -I OUTPUT -p tcp -m tcp --match multiport --dports 111,2049,20048 -j ACCEPT
Once you've determined that your hosts are using the standard ports and that you have opened those ports, you can start using SharedV4 volumes.
Manually assign and open NFS ports
For certain Linux distributions, the OS chooses the mountd
port randomly every time the node reboots. To solve this, you must manually assign NFS ports, and how you accomplish this depends on your OS.
Only perform the steps in one of the following sections if one of the following is true:
- The
mountd
port is not fixed (and not the standard port of 20048) and is chosen at random by your Linux distribution. - You wish to open a contiguous range of ports for Portworx and want to shift the default NFS ports to your Portworx port range.
In order to manually assign and open NFS ports, follow the steps in the section that applies for your OS.
Assign NFS ports on the RedHat family of Linux (RHEL, CentOS, Fedora, etc)
-
Modify the
/etc/sysconfig/nfs
file, uncommenting or adding the following fields and assigning the associated values:- LOCKD_TCPPORT=9023
- LOCKD_UDPPORT=9024
- MOUNTD_PORT=9025
- STATD_PORT=9026
-
Enter the following command to restart the NFS server:
systemctl restart nfs-server
-
Open the newly assigned NFS ports on your access control list:
iptables -I INPUT -p tcp -m tcp --match multiport --dports 111,2049,9023,9025,9026 -j ACCEPT
iptables -I OUTPUT -p tcp -m tcp --match multiport --dports 111,2049,9023,9025,9026 -j ACCEPT
iptables -I INPUT -p udp -m udp --dport 9024 -j ACCEPT
iptables -I OUTPUT -p udp -m udp --dport 9024 -j ACCEPT
Open NFS ports on Debian or Ubuntu Linux
Debian 10 / Ubuntu
-
Modify the
/etc/default/nfs-kernel-server
file. Uncomment or add theRPCMOUNTDOPTS
field and add the--port 9024
option to the value:...
RPCMOUNTDOPTS="--manage-gids --port 9024" -
Enter the following command to restart the
mountd
service:systemctl restart nfs-mountd.service
-
Verify that the
mountd
service is running on the port that you configured by searching the output ofrpcinfo -p
:rpcinfo -p | grep 'tcp.*mountd'
100005 1 tcp 9024 mountd
100005 2 tcp 9024 mountd
100005 3 tcp 9024 mountd
Debian 9 and lower
-
Modify the
/run/sysconfig/nfs-utils
file, uncommenting or adding the following fields and assigning the associated values:RPCNFSDARGS=" 8 --port 9023"
: append the--port 9023
option to any existing values.RPCMOUNTDARGS="--port 9024"
: add the--port 9024
option.STATDARGS="--port 9025 --outgoing-port 9026"
: add the--port 9025
and--outgoing-port 9026
options.
-
Enter the following commands to restart the NFS server:
systemctl daemon-reload
systemctl restart rpc-statd
systemctl restart rpc-mountd
systemctl restart nfs-server -
Open the newly assigned NFS ports on your access control list:
iptables -I INPUT -p tcp -m tcp --match multiport --dports 111,2049,9023,9025,9026 -j ACCEPT
iptables -I OUTPUT -p tcp -m tcp --match multiport --dports 111,2049,9023,9025,9026 -j ACCEPT
iptables -I INPUT -p udp -m udp --dport 9024 -j ACCEPT
iptables -I OUTPUT -p udp -m udp --dport 9024 -j ACCEPT
Open NFS ports on CoreOS
The following sharedv4 NFS services run on a node when Portworx is installed with sharedv4 support:
- portmapper
- status
- mountd
- nfs and nfs_acl
- nlockmgr
View these services and the ports they are using by entering the following command:
rpcinfo -p
By default, services like mountd
and nlockmgr
run on random ports and must be fixed to a specific port.
Configure nfs, nfs_acl, and lockd ports
By default, the nfs-server.service
configuration file is located under the following directory:
/usr/lib/systemd/system/nfs-server.service
-
Check the status of the existing
nfs-server.service
:systemctl status nfs-server
-
Copy the systemd unit file from the
usr
directory into theetc
directory:cp /usr/lib/systemd/system/nfs-server.service /etc/systemd/system/nfs-server.service
-
Open
/etc/systemd/system/nfs-server.service
in a text editor and, under the[Service]
section, add the--port 9023
value to theExecStart=/usr/sbin/rpc.nfsd
key:[Service]
Type=oneshot
RemainAfterExit=yes
ExecStartPre=/usr/sbin/exportfs -r
ExecStart=/usr/sbin/rpc.nfsd --port 9023
ExecStop=/usr/sbin/rpc.nfsd 0
ExecStopPost=/usr/sbin/exportfs -au
ExecStopPost=/usr/sbin/exportfs -f -
Update the
lockd
ports:echo 9027 > /proc/sys/fs/nfs/nlm_udpport
echo 9028 > /proc/sys/fs/nfs/nlm_tcpport -
Ensure that the
lockd
manager ports persist over node reboots by creating a100-nfs-ports.conf
file under the/etc/sysctl.d/
folder and adding the ports to it:cat /etc/sysctl.d/100-nfs-ports.conf
fs.nfs.nlm_tcpport = 9027
fs.nfs.nlm_udpport = 9028 -
Reload the
systemd
daemon and restart thenfs-server
service:systemctl daemon-reload
systemctl restart nfs-server -
Verify that the NFS services are running on the ports you configured by searching the output of
rpcinfo -p
:rpcinfo -p | grep nfs
100003 3 tcp 9023 nfs
100003 4 tcp 9023 nfs
100227 3 tcp 9023 nfs_aclrpcinfo -p | grep nlock
100021 1 udp 9027 nlockmgr
100021 3 udp 9027 nlockmgr
100021 4 udp 9027 nlockmgr
100021 1 tcp 9028 nlockmgr
100021 3 tcp 9028 nlockmgr
100021 4 tcp 9028 nlockmgr
Configure mountd services
-
Check the status of the existing
nfs-server.service
:systemctl status nfs-mountd
-
Copy the systemd unit file from
/usr
into/etc
:cp /usr/lib/systemd/system/nfs-mountd.service /etc/systemd/system/nfs-mountd.service
-
Open
/etc/systemd/system/nfs-mountd.service
in a text editor and, under the[Service]
section, add the--port 9024
value to theExecStart=/usr/sbin/rpc.mountd
key:[Unit]
...
[Service]
Type=forking
ExecStart=/usr/sbin/rpc.mountd --port 9024 -
Reload the
systemd
daemon and restart thenfs-server
service:systemctl daemon-reload
systemctl restart nfs-server -
Verify that the NFS services are running on the ports you configured by searching the output of
rpcinfo -p
:rpcinfo -p | grep mountd
100005 1 udp 9024 mountd
100005 1 tcp 9024 mountd
100005 2 udp 9024 mountd
100005 2 tcp 9024 mountd
100005 3 udp 9024 mountd
100005 3 tcp 9024 mountd
Configure statd services
-
Check the status of the existing
rpc-statd.service
:systemctl status rpc-statd
-
Copy the systemd unit file from the
usr
directory into theetc
directory:cp /usr/lib/systemd/system/rpc-statd.service /etc/systemd/system/rpc-statd.service
-
Open
/etc/systemd/system/rpc-statd.service
in a text editor and, under the[Service]
section, add the--port 9025
and--outgoing-port 9026
value to theExecStart=/usr/sbin/rpc.statd
key:[Unit]
...
[Service]
Environment=RPC_STATD_NO_NOTIFY=1
Type=forking
PIDFile=/var/run/rpc.statd.pid
ExecStart=/usr/sbin/rpc.statd --port 9025 --outgoing-port 9026 -
Reload the
systemd
daemon and restart therpc-statd
service:systemctl daemon-reload
systemctl restart rpc-statd -
Verify that the NFS services are running on the ports you configured by searching the output of
rpcinfo -p
:rpcinfo -p | grep status
100024 1 udp 9025 status
100024 1 tcp 9025 status
Open NFS ports on PhotonOS
-
Modify the
/etc/default/nfs-utils
file, adding parameters to assign ports to services:MOUNTD_OPTS="--port 9024"
NFSD_OPTS="--port 9023"
STATD_OPTS="--port 9025 --outgoing-port 9026" -
Add the
/etc/modprobe.d/lockd.conf
file to assign ports to thenlockmgr
service and reconfigure runtime ports:cat > /etc/modprobe.d/lockd.conf << .
options lockd nlm_tcpport=9027
options lockd nlm_udpport=9028
.
sysctl -w fs.nfs.nlm_tcpport=9027 fs.nfs.nlm_udpport=9028 -
Restart NFS services:
systemctl daemon-reload
systemctl restart nfs-server nfs-mountd rpc-statd -
Open up the ports 9001 to 9028 in the
iptables
firewall, then restart the service:# add Portworx ports 9001..9028 before "COMMIT"
sed -i~ '/^COMMIT/i -A INPUT -p tcp -m multiport --dports 111,9001:9028 -j ACCEPT' \
/etc/systemd/scripts/ip4save
systemctl restart iptables -
Verify that the NFS services are running on the ports that you just configured:
rpcinfo -p
program vers proto port service
100000 4 tcp 111 portmapper
100000 3 tcp 111 portmapper
100000 2 tcp 111 portmapper
100000 4 udp 111 portmapper
100000 3 udp 111 portmapper
100000 2 udp 111 portmapper
100005 1 udp 9024 mountd
100005 1 tcp 9024 mountd
100005 2 udp 9024 mountd
100005 2 tcp 9024 mountd
100005 3 udp 9024 mountd
100005 3 tcp 9024 mountd
100024 1 udp 9025 status
100024 1 tcp 9025 status
100003 3 tcp 9023 nfs
100003 4 tcp 9023 nfs
100021 1 udp 9027 nlockmgr
100021 3 udp 9027 nlockmgr
100021 4 udp 9027 nlockmgr
100021 1 tcp 9028 nlockmgr
100021 3 tcp 9028 nlockmgr
100021 4 tcp 9028 nlockmgr