DMF Downgrade/ Fresh Install
Please be advised downgrading of DMF version image using the “upgrade launch” procedure is not supported. There are 2 ways to perform the downgrades. Loading the alternate boot partition which consists of the previous DMF version image or doing a fresh install via IDRAC. IDRAC Fresh install may be used when you don’t have the previous DMF image on the alternate partition.
IDRAC Fresh Install
Using IDRAC we can perform a fresh install of the ISO on a device. The process is the same for Controller nodes, Analytics nodes, and Service nodes. Ensure you have the right ISO image ready for the device which you are preparing to perform a fresh install.
Please be advised:
- A fresh install via IDRAC may take some time.
- A fresh install via IDRAC will completely erase the configuration alongside the alternate partition.
- Save a copy of your configuration in a text format on a remote location to reconfigure your device after the fresh install.
- In a dual DMF controller environment, shutdown both DMF controllers and then perform the fresh install via IDRAC one after another.
To perform the IDRAC fresh install:
- Access the IDRAC web interface
- Click on the virtual console.
- Make sure you have the ISO ready on your local machine.
- Select virtual media in the top right corner.
- A drop-down will appear, select Map CD/DVD and browse for the ISO necessary
- Click map Device
- Closeout the page, once the device is mapped
- Select boot. (first button from the top right corner)
- Select virtual CD/DVD/ISO from Boot controls
- Select Reset system ( warm boot ). Do not power off or shut down the device.
Wait for the system to reboot (takes quite some time) and proceed with the normal installation procedure documented in the DMF-x.y.z-Deployment-Guide:
- First DMF controller -> Configuring the Active Controller Using the First Boot Script
- Second DMF controller -> Joining the Standby Controller to an Existing Cluster
- DMF service node -> Service Node Setup and Initial Configuration
- DMF recorder node -> DMF Recorder Installation Procedure
When reapplying the configuration on the DMF controller, the configuration saved in a text format must not be blindly copied and pasted, for two reasons:
- Text format of the configuration has passwords and keys in a hashed format that should be not be pasted.
- Everything in the “local node” submode are node specific and should not be pasted.
Once the fresh installation of the active DMF controller is complete, and configurations have been reapplied, reboot or power cycle all switches. The switches will automatically reinstall the SWL from the active DMF controller.
Booting with Alternate Partition
Please be advised that rolling back or downgrading an upgrade through the boot alternate partition is hit-full and will take several minutes to complete. To downgrade the system with the alternate partition image please follow the below steps,
Please be advised:
- Save a copy of your configuration in a text format on a remote location prior to booting the DMF controller from alternate partition.
- When booting from an alternate partition, the DMF controller will boot with whatever configuration existed prior to the upgrade.
- Any configuration changes made after the upgrade need to be manually added after rolling back.
1. Use the show boot partition command to verify the image present in the alternate partition on both the controllers.
RTPDMF-CTRL1-POD1# show boot partition State Upgrade Product Version Build -|------------|------------|-----------------------------|-------|-----| 1 Active, Boot completed DANZ Monitoring Fabric 8.0.1 15 2 completed DANZ Monitoring Fabric 8.0.0 62
RTPDMF-CTRL2-POD1# show boot partition State Upgrade Product Version Build -|------------|------------|-----------------------------|-------|-----| 1 Active, Boot completed DANZ Monitoring Fabric 8.0.1 15 2 completed DANZ Monitoring Fabric 8.0.0 62
2. Reboot the active controller node using the command #Boot partition alternate. Proceed with yes when prompted then repeat the same steps on the standby controller and allow some time for the controllers to reboot with the image on the alternate partition.
RTPDMF-CTRL2-POD1# boot partition alternate boot partition: reboot? ("y" or "yes" to continue): y Connection to 10.81.108.32 closed.
3. After both controllers have rebooted, verify the controllers are up and are clustered. Please verify it using the show controller command from both nodes.
4.Reboot your switches by entering system reboot switch all to reboot all of your switches to pull the right image from the controller.
RTPDMF-CTRL2-POD1(config)# system reboot switch all system switch reboot all: connected switches: mrv439-delivery mac: d4:af:f7:92:36:c4 mrv438-filter mac: d4:af:f7:92:91:bc egv368-core mac: d4:af:f7:ca:d3:7c reboot may cause service interruption system switch reboot all ("y" or "yes" to continue): y Requesting reboot: mrv439-delivery Requesting reboot: mrv438-filter Requesting reboot: egv368-core All Switches Rebooted: mrv439-delivery, mrv438-filter, egv368-core
RTPDMF-CTRL2-POD1(config)# show switch all
Please be advised that with certain DMF image versions, “system reboot switch all” may not reboot the switches.
- “system reboot switch all” can be executed prior to the step 2 if slightly longer downtime is acceptable.
- If the DMF controllers are already booted from the alternative partition and “system reboot switch all” is not rebooting all the switches, manually reboot or power cycle all the switches.
5. If there are DMF service nodes and/or DMF recorder nodes, these devices should now also be rebooted or power cycled.
6.Wait for the switches and service/recorder node to reconnect and verify the devices are running the same version as the controllers.
Created and reviewed by Kishore Jothinaryanan & Nahshon Blaine