|Question||How does "Centralized Image Upgrade" feature works on 6.3?|
|Environment||Master-Standby and locals|
In a muti-controller deployment this feature allows all the local/standby controllers associated to the master to upgrade their images. The feature allows the specific Local controller to be upgraded based on host ip/subnet and all controller who are associated to the master.
Configuration can done only via master which in-turn pushes the config to the all associated local/standby controllers when “write memory” is executed via the master.
Once the upgrade is initiated from the master; it will first verify the server to check the image header, version, build & type present is valid for all the associated controllers.
After the master verification is complete it will instruct all the associated locals/Standby to download the image present on the servers.
Following information would be passed by master to its locals after image verification; Image filename, serverip , protocol and credentials in case of ftp/scp
The local controller contacts the server directly in order to download/upgrade the images.
When there is a upgrade failure the local/standby controllers will attempt thrice to download the image file from the servers in case of any failure.
When it fails after 3 retries the controller will again attempt in the next 15 minutes in order to download the image and this process continues until the target IP is cleared or purged from the master controller
If the upgrade is successful then the local/standby controller will reboot if the option “autoboot” is enabled on the upgrade-profile
In case of any failure encountered by the local/standby controller then an error message will be logged in both local/standby and master controllers.
Flow chart on the sequence