• DMF Upgrade procedures 7.x.x+

 
 
Print Friendly, PDF & Email

DMF Upgrade Procedures

Overview

This Document will provide the necessary steps to upgrade DMF and its respected nodes individually from 7.x.x onward. These upgrade steps are also present in the DMF deployment guide, starting on chapter 6. https://www.arista.com/en/support/software-download

Switch/Node upgrades

After the controllers have upgraded, the switches will automatically reboot to upgrade. Service nodes, starting from 7.x.x, will upgrade automatically similarly to fabric switches. Analytics and Recorder nodes need to be upgraded separately, which will be covered later in this document.

 

Steps for Upgrading Controllers 7.x.x to 8.x.x

1. Download your intended upgrade version ISO from the software downloads page to your local device. https://www.arista.com/en/support/software-download

Copying Upgrade image to controller

(ISO images need to be copied to both the controllers individually)

Option 1

scp file direct to image:// partition

From the remote server, scp image file to the controller. 

# scp controller-bcf-5.0.0.iso admin@10.8.28.15://image

Or From the BMF controller, copy the image from the remote server

# copy scp://<userid>@<server_ip>:<iso_file_location> image://

Option 2

scp file to /home/admin and to image://
From remote server, scp image file to the controller’s /home/admin directory.

controller# scp BMF-7.1.1-Controller-Appliance-2019-05-16.iso admin@10.8.28.15:BMF-7.1.1-Controller-Appliance-2019-05-16.iso

Then copy the file from /home/admin directory to image:// partition.

controller>enable 
controller# copy file://BMF-7.1.1-Controller-Appliance-2019-05-16.iso image://
controller# show image

Controller upgrade

2. Log in to both controllers with their individual IP addresses, not the Virtual IP address of the cluster.

3. Ensure the ISOs are in the image repository of both controllers. Using Show Image command

Controller> Show image

Image Product                Version Build

-|-----|----------------------|-------|-----|

1 43e61 DANZ Monitoring Fabric 8.0.1   15

2 b2dc0 DANZ Monitoring Fabric 8.0.0   62

 

4. Run the Upgrade Stage command on the both controllers.

This copies the running-config to a safe location and marks the alternate partition as the boot partition. Proceed with yes when prompted to continue with staging the upgrade.

RTPDMF-CTRL2-POD1# upgrade stage

<Image Id>   <string>  (optional)

43e61      Image: BMF 8.0.1 15

b2dc0      Image: BMF 8.0.0 62

RTPDMF-CTRL2-POD1# upgrade stage 43e61

Upgrade stage will overwrite alternate partition, proceed ("y" or "yes" to continue): y

Upgrade stage: deleted upgrader account

Verifying the integrity of the installation media

Staging the upgrade to DANZ Monitoring Fabric 8.0.1 (dmf-8.0.1 #15)

5. Verify the image has been set to the  alternate boot partition on both controllers using #Show boot partition.

RTPDMF-CTRL2-POD1# show boot partition

      State            Upgrade                     Product                  Version      Build

-|--------------|----------|----------------------------|---------|-------|

1                             staged        DANZ Monitoring Fabric       8.0.1      15

2 Active,Boot      completed    DANZ Monitoring Fabric     8.0.0      62


6. Proceed to launch the upgrade, using the command upgrade launch. Run upgrade launch on the active controller first and select yes when prompted. Next, run upgrade launch on the standby controller.

Note: the active controller will not proceed with the upgrade until the upgrade launch command has been initiated on the standby controller.

 

RTPDMF-CTRL2-POD1# upgrade launch

Upgrade launch: DANZ Monitoring Fabric 8.0.1 (dmf-8.0.1 #15)

Upgrade launch: Various cluster members and managed devices may be rebooted by automation

Upgrade launch: proceed? ("y" or "yes" to continue): y

Upgrade launch: Various cluster members may be rebooted by automation

Upgrade launch: 18:17:46: Starting Upgrade

Upgrade launch: 18:17:46: waiting for standby to begin "upgrade launch"

7. During this time, the standby controller will reboot and upgrade first. Once the standby comes back up with the new version, it will become the active controller. The former active controller will wait for the new active controller to become functional, and then will commence its upgrade and join the cluster as standby.

8. The switches and service nodes will reload automatically following the controllers. Service nodes could take up to 30 minutes especially if you have multiple nodes. After some time please verify the switches and service nodes are connected back to the fabric.

controller# show switch all

Switch Name     IP Address   State     Pipeline Mode

-|---------------|------------|---------|---------------------|

1 egv368-core     10.81.108.39 connected l3-l4-match-push-vlan

2 mrv438-filter   10.81.108.38 connected l3-l4-match-push-vlan

3 mrv439-delivery 10.81.108.40 connected l3-l4-match-push-vlan

controller# show service-node

Service Node     Switch MAC Address       Connected Since                IP Address

-|----------------|------------------------|------------------------------|------------|

1 service-node-rtp e4:43:4b:a8:b0:ca (Dell) 2021-03-11 18:45:55.448000 UTC 10.81.108.36

 

Analytics and Recorder Nodes

The Analytics and Recorder node must be upgraded individually. If the service nodes need to be upgraded manually, the same procedures described in this section can be used.

For this example, we are upgrading a Recorder node.

1. Move the running configurations from the nodes to a remote server as a .txt file.

2. Download the node ISO from the software downloads page.(Please ensure the downloaded iso is for the correct node. The name of the node should be present in its respected iso file .DMF-8.0.1-Recorder-Node-Appliance-2021-03-01.iso )

3. SCP or copy the correct ISO to the Node, using the same steps mentioned earlier for the controllers.

4.Verify the images are present in the image repository; using the Show image command

Packet-Recorder-RTP# show image

Image Product           Version Build

-|-----|-----------------|-------|-----|

1 e102c DMF Recorder Node 8.0.1   11

5. Commence Upgrade Stage command on the device and proceed when asked to continue.

Packet-Recorder-RTP# upgrade stage

Upgrade stage will overwrite alternate partition, proceed ("y" or "yes" to continue): y

Verifying the integrity of the installation media

Staging the upgrade to DMF Recorder Node 8.0.1 (recorder-node/dmf-8.0.1 #11)

00:01:33: progress:  89% |********************************************************->    

6. Verify the upgrade is staged using show boot partition command on the Packet-recorder.


Packet-Recorder-RTP# show boot partition

    State        Upgrade          Product           Version Build

-|------------|--------|-----------------|-------|-----|

1 Active, Boot original DMF Recorder Node 8.0.0   58

2                       staged   DMF Recorder Node 8.0.1   11

7. commence upgrade launch command

Packet-Recorder-RTP# upgrade launch

Upgrade launch: DMF Recorder Node 8.0.1 (recorder-node/dmf-8.0.1 #11)

Upgrade launch: Various cluster members may be rebooted by automation

Upgrade launch: proceed? ("y" or "yes" to continue): y

Upgrade launch: WARNING single-controller: upgrade will be non-redundant

Upgrade launch: non-redundant  upgrade ("y" or "yes" to continue): y

00:00:03: progress:   0% |->      

8. Allow time for the node to upgrade and reboot.verify the software version once its boots back up.

nblaine@nblaine ~ % ssh admin@10.81.108.35

DMF Recorder Node 8.0.1 (recorder-node/dmf-8.0.1 #11)

 

 

Created and revised by: Nahshon Blaine and Kishore Jothinarayanan

Follow

Get every new post on this blog delivered to your Inbox.

Join other followers: