Configure the NVIDIA IP SN2100 switch
You must configure the NVIDIA SN2100 IP switches for use as the cluster interconnect and for backend MetroCluster IP connectivity.
Reset the NVIDIA IP SN2100 switch to factory defaults
You can choose from the following methods to reset a switch to factory default settings.
Reset the switch using the RCF file option
Before installing a new RCF configuration you must revert the NVIDIA switch settings.
To restore the switch to default settings, run the RCF file with the restoreDefaults
option. This option copies the original backed up files to their original location and then reboots the switch. After reboot, the switch comes online with the original configuration that existed when you first ran the RCF file to configure the switch.
The following configuration details are not reset:
-
User and credential configuration
-
Configuration of the management network port, eth0
All other configuration changes that occur during application of the RCF file are reverted to the original configuration. |
-
You must configure the switch according to Download and install the NVIDIA RCF file. If you have not configured in this manner, or you have configured additional features before running the RCF file, you cannot use this procedure.
-
You must repeat these steps on each of the IP switches in the MetroCluster IP configuration.
-
You must be connected to the switch with a serial console connection.
-
This task resets the configuration of the management network.
-
Verify that the RCF configuration was successfully applied with the same or a compatible RCF file version and that the backup files exist.
The output can show backup files, preserved files, or both. If backup files or preserved files do not appear in the output, you cannot use this procedure. cumulus@IP_switch_A_1:mgmt:~$ sudo python3 SN2100_v2.0.0_IP_switch_A_1.py [sudo] password for cumulus: >>> Opened RcfApplyLog A RCF configuration has been successfully applied. Backup files exist. Preserved files exist. Listing completion of the steps: Success: Step: 1: Performing Backup and Restore Success: Step: 2: updating MOTD file Success: Step: 3: Disabling apt-get Success: Step: 4: Disabling cdp Success: Step: 5: Adding lldp config Success: Step: 6: Creating interfaces Success: Step: 7: Configuring switch basic settings: Hostname, SNMP Success: Step: 8: Configuring switch basic settings: bandwidth allocation Success: Step: 9: Configuring switch basic settings: ecn Success: Step: 10: Configuring switch basic settings: cos and dscp remark Success: Step: 11: Configuring switch basic settings: generic egress cos mappings Success: Step: 12: Configuring switch basic settings: traffic classification Success: Step: 13: Configuring LAG load balancing policies Success: Step: 14: Configuring the VLAN bridge Success: Step: 15: Configuring local cluster ISL ports Success: Step: 16: Configuring MetroCluster ISL ports Success: Step: 17: Configuring ports for MetroCluster-1, local cluster and MetroCluster interfaces Success: Step: 18: Configuring ports for MetroCluster-2, local cluster and MetroCluster interfaces Success: Step: 19: Configuring ports for MetroCluster-3, local cluster and MetroCluster interfaces Success: Step: 20: Configuring L2FC for MetroCluster interfaces Success: Step: 21: Configuring the interface to UP Success: Step: 22: Final commit Success: Step: 23: Final reboot of the switch Exiting ... <<< Closing RcfApplyLog cumulus@IP_switch_A_1:mgmt:~$
-
Run the RCF file with the option to restore defaults:
restoreDefaults
cumulus@IP_switch_A_1:mgmt:~$ sudo python3 SN2100_v2.0.0_IP_switch_A_2.py restoreDefaults [sudo] password for cumulus: >>> Opened RcfApplyLog Can restore from backup directory. Continuing. This will reboot the switch !!! Enter yes or no: yes
-
Respond 'yes' to the prompt. The switch reverts to the original configuration and reboots.
-
Wait for the switch to reboot.
The switch is reset and retains the initial configuration such as management network configuration and current credentials as they existed before applying the RCF file. After reboot, you can apply a new configuration by using the same or a different version of the RCF file.
Download and install the Cumulus software
Use these steps if you want to reset the switch completely by applying the Cumulus image.
-
You must be connected to the switch with a serial console connection.
-
The Cumulus switch software image is accessible through HTTP.
For more information about installing Cumulus Linux, see Overview of installation and configuration for NVIDIA SN2100 switches -
You must have the root password for
sudo
access to the commands.
-
From the Cumulus console download and queue the switch software installation with the command
onie-install -a -i
followed by the file path to the switch software:In this example, the firmware file
cumulus-linux-4.4.3-mlx-amd64.bin
is copied from the HTTP server '50.50.50.50' to the local switch.cumulus@IP_switch_A_1:mgmt:~$ sudo onie-install -a -i http://50.50.50.50/switchsoftware/cumulus-linux-4.4.3-mlx-amd64.bin Fetching installer: http://50.50.50.50/switchsoftware/cumulus-linux-4.4.3-mlx-amd64.bin Downloading URL: http://50.50.50.50/switchsoftware/cumulus-linux-4.4.3-mlx-amd64.bin ######################################################################### 100.0% Success: HTTP download complete. tar: ./sysroot.tar: time stamp 2021-01-30 17:00:58 is 53895092.604407122 s in the future tar: ./kernel: time stamp 2021-01-30 17:00:58 is 53895092.582826352 s in the future tar: ./initrd: time stamp 2021-01-30 17:00:58 is 53895092.509682557 s in the future tar: ./embedded-installer/bootloader/grub: time stamp 2020-12-10 15:25:16 is 49482950.509433937 s in the future tar: ./embedded-installer/bootloader/init: time stamp 2020-12-10 15:25:16 is 49482950.509336507 s in the future tar: ./embedded-installer/bootloader/uboot: time stamp 2020-12-10 15:25:16 is 49482950.509213637 s in the future tar: ./embedded-installer/bootloader: time stamp 2020-12-10 15:25:16 is 49482950.509153787 s in the future tar: ./embedded-installer/lib/init: time stamp 2020-12-10 15:25:16 is 49482950.509064547 s in the future tar: ./embedded-installer/lib/logging: time stamp 2020-12-10 15:25:16 is 49482950.508997777 s in the future tar: ./embedded-installer/lib/platform: time stamp 2020-12-10 15:25:16 is 49482950.508913317 s in the future tar: ./embedded-installer/lib/utility: time stamp 2020-12-10 15:25:16 is 49482950.508847367 s in the future tar: ./embedded-installer/lib/check-onie: time stamp 2020-12-10 15:25:16 is 49482950.508761477 s in the future tar: ./embedded-installer/lib: time stamp 2020-12-10 15:25:47 is 49482981.508710647 s in the future tar: ./embedded-installer/storage/blk: time stamp 2020-12-10 15:25:16 is 49482950.508631277 s in the future tar: ./embedded-installer/storage/gpt: time stamp 2020-12-10 15:25:16 is 49482950.508523097 s in the future tar: ./embedded-installer/storage/init: time stamp 2020-12-10 15:25:16 is 49482950.508437507 s in the future tar: ./embedded-installer/storage/mbr: time stamp 2020-12-10 15:25:16 is 49482950.508371177 s in the future tar: ./embedded-installer/storage/mtd: time stamp 2020-12-10 15:25:16 is 49482950.508293856 s in the future tar: ./embedded-installer/storage: time stamp 2020-12-10 15:25:16 is 49482950.508243666 s in the future tar: ./embedded-installer/platforms.db: time stamp 2020-12-10 15:25:16 is 49482950.508179456 s in the future tar: ./embedded-installer/install: time stamp 2020-12-10 15:25:47 is 49482981.508094606 s in the future tar: ./embedded-installer: time stamp 2020-12-10 15:25:47 is 49482981.508044066 s in the future tar: ./control: time stamp 2021-01-30 17:00:58 is 53895092.507984316 s in the future tar: .: time stamp 2021-01-30 17:00:58 is 53895092.507920196 s in the future Staging installer image...done. WARNING: WARNING: Activating staged installer requested. WARNING: This action will wipe out all system data. WARNING: Make sure to back up your data. WARNING: Are you sure (y/N)? y Activating staged installer...done. Reboot required to take effect. cumulus@IP_switch_A_1:mgmt:~$
-
Respond
y
to the prompt to confirm the installation when the image is downloaded and verified. -
Reboot the switch to install the new software:
sudo reboot
cumulus@IP_switch_A_1:mgmt:~$ sudo reboot
The switch reboots and enters the switch software installation which takes some time. When the installation is complete, the switch reboots and remains at the 'log-in' prompt. -
Configure the basic switch settings
-
When the switch is booted and at the log-in prompt, log in and change the password.
The username is 'cumulus' and the default password is 'cumulus'.
Debian GNU/Linux 10 cumulus ttyS0 cumulus login: cumulus Password: You are required to change your password immediately (administrator enforced) Changing password for cumulus. Current password: New password: Retype new password: Linux cumulus 4.19.0-cl-1-amd64 #1 SMP Cumulus 4.19.206-1+cl4.4.3u1 (2021-12-18) x86_64 Welcome to NVIDIA Cumulus (R) Linux (R) For support and online technical documentation, visit http://www.cumulusnetworks.com/support The registered trademark Linux (R) is used pursuant to a sublicense from LMI, the exclusive licensee of Linus Torvalds, owner of the mark on a world-wide basis. cumulus@cumulus:mgmt:~$
-
-
Configure the management network interface.
The commands you use depend on the switch firmware version you are running.
The following example commands configure the hostname as IP_switch_A_1, the IP address as 10.10.10.10, the netmask as 255.255.255.0 (24), and the gateway address as 10.10.10.1. Cumulus 4.4.xThe following example commands configure the hostname, IP address, netmask, and gateway on a switch running Cumulus 4.4.x.
cumulus@cumulus:mgmt:~$ net add hostname IP_switch_A_1 cumulus@cumulus:mgmt:~$ net add interface eth0 ip address 10.0.10.10/24 cumulus@cumulus:mgmt:~$ net add interface eth0 ip gateway 10.10.10.1 cumulus@cumulus:mgmt:~$ net pending . . . cumulus@cumulus:mgmt:~$ net commit . . . net add/del commands since the last "net commit" User Timestamp Command cumulus 2021-05-17 22:21:57.437099 net add hostname Switch-A-1 cumulus 2021-05-17 22:21:57.538639 net add interface eth0 ip address 10.10.10.10/24 cumulus 2021-05-17 22:21:57.635729 net add interface eth0 ip gateway 10.10.10.1 cumulus@cumulus:mgmt:~$
Cumulus 5.4.x and laterThe following example commands configure the hostname, IP address, netmask, and gateway on a switch running Cumulus 5.4.x. or later.
cumulus@cumulus:mgmt:~$ nv set system hostname IP_switch_A_1 cumulus@cumulus:mgmt:~$ nv set interface eth0 ip address 10.0.10.10/24 cumulus@cumulus:mgmt:~$ nv set interface eth0 ip gateway 10.10.10.1 cumulus@cumulus:mgmt:~$ nv config apply cumulus@cumulus:mgmt:~$ nv config save
-
Reboot the switch using the
sudo reboot
command.cumulus@cumulus:~$ sudo reboot
When the switch reboots, you can apply a new configuration using the steps in Download and install the NVIDIA RCF file.
Download and install the NVIDIA RCF files
You must generate and install the switch RCF file to each switch in the MetroCluster IP configuration.
-
You must have the root password for
sudo
access to the commands. -
The switch software is installed and the management network is configured.
-
You followed the steps to initially install the switch by using either method 1 or method 2.
-
You did not apply any additional configuration after the initial installation.
If you perform further configuration after resetting the switch and before applying the RCF file, you cannot use this procedure.
You must repeat these steps on each of the IP switches in the MetroCluster IP configuration (new installation) or on the replacement switch (switch replacement).
-
Generate the NVIDIA RCF files for MetroCluster IP.
-
Download the RcfFileGenerator for MetroCluster IP.
-
Generate the RCF file for your configuration by using the RcfFileGenerator for MetroCluster IP.
-
Navigate to your home directory. If you are logged as 'cumulus', the file path is
/home/cumulus
.cumulus@IP_switch_A_1:mgmt:~$ cd ~ cumulus@IP_switch_A_1:mgmt:~$ pwd /home/cumulus cumulus@IP_switch_A_1:mgmt:~$
-
Download the RCF file to this directory. The following example shows that you use SCP to download the file
SN2100_v2.0.0_IP_switch_A_1.txt
from server '50.50.50.50' to your home directory and save it asSN2100_v2.0.0_IP_switch_A_1.py
:cumulus@Switch-A-1:mgmt:~$ scp username@50.50.50.50:/RcfFiles/SN2100_v2.0.0_IP_switch_A_1.txt ./SN2100_v2.0.0_IP_switch-A1.py The authenticity of host '50.50.50.50 (50.50.50.50)' can't be established. RSA key fingerprint is SHA256:B5gBtOmNZvdKiY+dPhh8=ZK9DaKG7g6sv+2gFlGVF8E. Are you sure you want to continue connecting (yes/no)? yes Warning: Permanently added '50.50.50.50' (RSA) to the list of known hosts. *********************************************************************** Banner of the SCP server *********************************************************************** username@50.50.50.50's password: SN2100_v2.0.0_IP_switch_A1.txt 100% 55KB 1.4MB/s 00:00 cumulus@IP_switch_A_1:mgmt:~$
-
-
Execute the RCF file. The RCF file requires an option to apply one or more steps. Unless instructed by technical support, run the RCF file without the command line option. To verify the completion status of the various steps of the RCF file, use the option '-1' or 'all' to apply all (pending) steps.
cumulus@IP_switch_A_1:mgmt:~$ sudo python3 SN2100_v2.0.0_IP_switch_A_1.py all [sudo] password for cumulus: The switch will be rebooted after the step(s) have been run. Enter yes or no: yes ... the steps will apply - this is generating a lot of output ... Running Step 24: Final reboot of the switch ... The switch will reboot if all steps applied successfully ...
-
If your configuration uses DAC cables, enable the DAC option on the switch ports:
cumulus@IP_switch_A_1:mgmt:~$ sudo python3 SN2100_v2.0.0-X10_Switch-A1.py runCmd <switchport> DacOption [enable | disable]
The following example enables the DAC option for port
swp7
:cumulus@IP_switch_A_1:mgmt:~$ sudo python3 SN2100_v2.00_Switch-A1.py runCmd swp7 DacOption enable Running cumulus version : 5.4.0 Running RCF file version : v2.00 Running command: Enabling the DacOption for port swp7 runCmd: 'nv set interface swp7 link fast-linkup on', ret: 0 runCmd: committed, ret: 0 Completion: SUCCESS cumulus@IP_switch_A_1:mgmt:~$
-
Reboot the switch after enabling the DAC option on the switch ports:
sudo reboot
When you set the DAC option for multiple switch ports, you only need to reboot the switch once.
Set the switch port speed for the MetroCluster IP interfaces
Use this procedure to set the switch port speed to 100G for the following systems:
-
AFF A70
-
AFF A90
-
AFF A1K
-
Use the RCF file with the
runCmd
option to set the speed. This applies the setting and saves the configuration.The following commands set the speed for the MetroCluster interfaces
swp7
andswp8
:sudo python3 SN2100_v2.20 _Switch-A1.py runCmd swp7 speed 100
sudo python3 SN2100_v2.20 _Switch-A1.py runCmd swp8 speed 100
Example
cumulus@Switch-A-1:mgmt:~$ sudo python3 SN2100_v2.20_Switch-A1.py runCmd swp7 speed 100 [sudo] password for cumulus: <password> Running cumulus version : 5.4.0 Running RCF file version : v2.20 Running command: Setting switchport swp7 to 100G speed runCmd: 'nv set interface swp7 link auto-negotiate off', ret: 0 runCmd: 'nv set interface swp7 link speed 100G', ret: 0 runCmd: committed, ret: 0 Completion: SUCCESS cumulus@Switch-A-1:mgmt:~$
Disable unused ISL ports and port channels
NetApp recommends disabling unused ISL ports and port channels to avoid unnecessary health alerts.
-
Identify the unused ISL ports and port channels using the RCF file banner:
If the port is in breakout mode, the port name you specify in the command might be different than the name stated in the RCF banner. You can also use the RCF cabling files to find the port name. net show interface
-
Disable the unused ISL ports and port channels using the RCF file.
cumulus@mcc1-integrity-a1:mgmt:~$ sudo python3 SN2100_v2.0_IP_Switch-A1.py runCmd [sudo] password for cumulus: Running cumulus version : 5.4.0 Running RCF file version : v2.0 Help for runCmd: To run a command execute the RCF script as follows: sudo python3 <script> runCmd <option-1> <option-2> <option-x> Depending on the command more or less options are required. Example to 'up' port 'swp1' sudo python3 SN2100_v2.0_IP_Switch-A1.py runCmd swp1 up Available commands: UP / DOWN the switchport sudo python3 SN2100_v2.0_IP_Switch-A1.py runCmd <switchport> state <up | down> Set the switch port speed sudo python3 SN2100_v2.0_Switch-A1.py runCmd <switchport> speed <10 | 25 | 40 | 100 | AN> Set the fec mode on the switch port sudo python3 SN2100_v2.0_Switch-A1.py runCmd <switchport> fec <default | auto | rs | baser | off> Set the [localISL | remoteISL] to 'UP' or 'DOWN' state sudo python3 SN2100_v2.0_Switch-A1.py runCmd [localISL | remoteISL] state [up | down] Set the option on the port to support DAC cables. This option does not support port ranges. You must reload the switch after changing this option for the required ports. This will disrupt traffic. This setting requires Cumulus 5.4 or a later 5.x release. sudo python3 SN2100_v2.0_Switch-A1.py runCmd <switchport> DacOption [enable | disable] cumulus@mcc1-integrity-a1:mgmt:~$
The following example command disables port "swp14":
sudo python3 SN2100_v2.0_Switch-A1.py runCmd swp14 state down
Repeat this step for each identified unused port or port channel.