As I see, only the port sharing config (Link Aggregation) was lost. On every node the 10g Dual License was installed and on every node Port 51 was part of Port sharing with SR sfp+ optic. Port sharing Master was on Stack Master node. For me there is NO reason for Switch Stack to change roles when the hole Stack is rebooted at once. In all 3 cases the config was saved and synchroniced. As workaround I Set priority in Stack. But I think this couldn't be normal.
Peter, I would recommend verifying how the LAG is configured. Is the LAG static or Dynamic? A show sharing output from prior and during reboot can provide some information on the active links. A dynamic setup would be recommended on both sides of the LAG.
Hello Peter, could you be more specific on what exactly is lost? Are you referring to the port sharing configuration or the entire configuration?
If it is related to port sharing, please let us know which slots the uplink ports reside on. If port sharing or the Link Aggregation Group is connected to multiple slots then there should be redundancy when it comes to port sharing or Link Aggregation Groups.
Roles will change from Master to Backup during boot, so this would be expected behaviour when rebooting one slot and not multiple.
Please also ensure that the stack is synchronized to ensure the configuration is stored on bothe the backup and master nodes.