In a HA setup, the master/active and all slave/standby nodes must have exactly the same set of fxs files loaded. It sounds like you are doing the upgrade by restarting the standby ConfD with the new set of fxs files, while the active ConfD is still running with the old set - this is fine per se, but you can’t at that point tell the standby to connect to the active via
confd_ha_beslave() - it will result in errors like the one you show being logged by the active, while
confd_ha_beslave() fails with this
confd_errno value (from
A remote HA node had a different set of fxs files compared to us.
It could also be that the set is the same, but the version of some
fxs file is different.
Of course your HAFW “ought to” have detected this.
Upgrade and HA in the
In-service Data Model Upgrade chapter in the User Guide describes how to do a “rolling” upgrade in a HA setup. While the description is primarily targeted for the In-Service upgrade case (where the upgrade will fail if the procedure isn’t followed), the same general principle applies when upgrading via ConfD restart, as pointed out by the last paragraph in that section.