Prepare the worker node
All of the worker nodes in the Kubernetes cluster need to be able to mount the volumes that you have provisioned for your pods. If you are using the ontap-nas
, ontap-nas-economy
, or ontap-nas-flexgroup
driver for one of your backends, your worker nodes need the NFS tools. Otherwise they require the iSCSI tools.
Recent versions of RedHat CoreOS have both NFS and iSCSI installed by default.
You should always reboot your worker nodes after installing the NFS or iSCSI tools, or else attaching volumes to containers might fail. |
NFS volumes
Protocol | Operating system | Commands |
---|---|---|
NFS |
RHEL/CentOS |
|
NFS |
Ubuntu/Debian |
|
You should ensure that the NFS service is started up during boot time. |
iSCSI volumes
Consider the following when using iSCSI volumes:
-
Each node in the Kubernetes cluster must have a unique IQN. This is a necessary prerequisite.
-
If using RHCOS version 4.5 or later, or RHEL or CentOS version 8.2 or later with the
solidfire-san
driver, ensure that the CHAP authentication algorithm is set to MD5 in/etc/iscsi/iscsid.conf
.
sudo sed -i 's/^\(node.session.auth.chap_algs\).*/\1 = MD5/' /etc/iscsi/iscsid.conf
-
When using worker nodes that run RHEL/RedHat CoreOS with iSCSI PVs, make sure to specify the
discard
mountOption in the StorageClass to perform inline space reclamation. See RedHat’s documentation.
Protocol | Operating system | Commands | ||
---|---|---|---|---|
iSCSI |
RHEL/CentOS |
|
||
iSCSI |
Ubuntu/Debian |
|
For Ubuntu 18.04, you must discover target ports with iscsiadm before starting open-iscsi for the iSCSI daemon to start. You can alternatively modify the iscsi service to start iscsid automatically.
|
If you want to learn more about automatic worker node preparation, which is a beta feature, see here. |