From 0f2fba46ad4f5bcce6ad72098f51bf464eac913c Mon Sep 17 00:00:00 2001 From: Aram Sargsyan Date: Tue, 15 Apr 2025 14:57:52 +0000 Subject: [PATCH] Document the new unreachable cache behavior Update the documentaion to include information about how the cache's exponential backoff works, and how to clear the cache. --- doc/arm/config-auth.inc.rst | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/doc/arm/config-auth.inc.rst b/doc/arm/config-auth.inc.rst index b054fb0904..77d1c7ae99 100644 --- a/doc/arm/config-auth.inc.rst +++ b/doc/arm/config-auth.inc.rst @@ -87,6 +87,13 @@ The numbers in parentheses in the following text refer to the numbered items in from the primary (as described in section 2 a. above). If the zone file has changed, propagation is practically immediate. +.. Note:: When a primary server is unreachable, :iscman:`named` initially caches + that information for 10 seconds. After that initial period, :iscman:`named` + may try that server again; if the server remains unresponsive, + :iscman:`named` caches its unreachable status for up to 640 seconds using an + exponential backoff. During that time :iscman:`named` does not try to contact + the affected server. The cache can be cleared using :option:`rndc flush`. + The authoritative samples all use NOTIFY but identify the statements used, so that they can be removed if not required.