DescriptionThis article describes how to proceed when the Slave node is stuck in reload loop. This problem can occasionally happen after upgrade and is often caused by incorrect (impossible) configuration on the Master unit.
SolutionThis can usually be solved by disabling HA to stop the reloading of the Slave unit, and then swapping roles, waiting for sync and then reverting to the original roles.
1) On Master unit
Stop reloading of the slave by running:
config system ha
set mode off
end
2) Change roles
On Slave unit (unit B)
config system ha
set mode master
end
On Master unit (unit A)
config system ha
set mode slave
end
3) Wait for synchronization (unit A may reload once), then swap back to the original roles
On unit B
config system ha
set mode slave
end
On unit A
config system ha
set mode master
end
To verify that everything is working correctly, do a config change on the master unit and check on the slave unit that it was synchronized correctly.
For example, create dummy session profile.
Master # config profile session
edit test-profile
end
Then verify that the setting was correctly transferred to the slave unit
Master # get profile session
== [ session_basic_predefined ]
== [ Outbound_Session ]
== [ Inbound_Session ]
== [ test-profile ]
Slave # get profile session
== [ session_basic_predefined ]
== [ Outbound_Session ]
== [ Inbound_Session ]
== [ test-profile ]
In case switching the roles would not fix the issue (slave unit would still keep reloading), create a ticket with support and provide following information:
* configuration backup from both units
* output of following, also from both units
diag sys ha show
diag sys ha show global