When refreshing a zone from primaries, if a XoT configured primary is unreachable, the secondary doesn't try the next one
Summary
It looks like something broke the 'fallback to the next primary' when a secondary is processing a notify and the first primary in the list is unreachable.
BIND logs the usual: got_transfer_quota: skipping zone transfer as primary ****** (source ********) is unreachable (cached)
But then instead of going off to try the next primary, just gives up.
This works fine on 9.18
(edit by @aram: see addendum at the bottom)
BIND version affected
Test version of 9.20 was 9.20.2 Test version of 9.18 was 9.18.28
More details can be found in Support Ticket #24186
Steps to reproduce
What is the current bug behavior?
The first primary tried fails, and the zone refresh stops and doesn't try the next primary
What is the expected correct behavior?
The first primary being unavailable shouldn't matter if the second one listed is there - the secondary should fall back and transfer from that one.
Relevant configuration files
See Support Ticket #24186 (but this is not difficult to set up and test)
Relevant logs
See also Support Ticket #24186
Addendum by @aram
I believe this affects only XoT, so the title was edited to convey that.