From 70fbbedc247ea94d55c5b73e1fd2b2a907624a28 Mon Sep 17 00:00:00 2001
From: Matthijs Mekking <matthijs@isc.org>
Date: Thu, 4 Feb 2021 15:07:40 +0100
Subject: [PATCH] Adjust serve-stale test

The number of queries to use in the burst can be reduced, as we have
a very low fetch limit of 1.

The dig command in 'wait_for_fetchlimits()' should time out sooner as
we expect a SERVFAIL to be returned promptly.

Enabling serve-stale can be done before hitting fetch-limits. This
reduces the chance that the resolver queries time out and fetch count
is reset. The chance of that happening is already slim because
'resolver-query-timeout' is 10 seconds, but better to first let the
data become stale rather than doing that while attempting to resolve.

(cherry picked from commit 00f575e7ef8c94f5894fe8b669ef3817af8a105f)
---
 bin/tests/system/serve-stale/tests.sh | 60 ++++++++++++++++-----------
 1 file changed, 35 insertions(+), 25 deletions(-)

diff --git a/bin/tests/system/serve-stale/tests.sh b/bin/tests/system/serve-stale/tests.sh
index 72af34037c5..532473544bd 100755
--- a/bin/tests/system/serve-stale/tests.sh
+++ b/bin/tests/system/serve-stale/tests.sh
@@ -2016,6 +2016,9 @@ status=$((status+ret))
 ####################################################################
 echo_i "test stale data with fetch-limits"
 
+# We update the named configuration to enable fetch-limits. The fetch-limits
+# are set to 1, which is ridiciously low, but that is because for this test we
+# want to reach the fetch-limits.
 n=$((n+1))
 echo_i "updating ns3/named.conf ($n)"
 ret=0
@@ -2030,7 +2033,8 @@ rndc_reload ns3 10.53.0.3
 if [ $ret != 0 ]; then echo_i "failed"; fi
 status=$((status+ret))
 
-# Disable responses from authoritative server.
+# Disable responses from authoritative server. If we can't resolve the example
+# zone, fetch limits will be reached.
 n=$((n+1))
 echo_i "disable responses from authoritative server ($n)"
 ret=0
@@ -2040,21 +2044,44 @@ grep "TXT.\"0\"" dig.out.test$n > /dev/null || ret=1
 if [ $ret != 0 ]; then echo_i "failed"; fi
 status=$((status+ret))
 
-# Hit the fetch-limits.
+# Allow RRset to become stale.
+sleep 2
+
+# Turn on serve-stale.
+n=$((n+1))
+echo_i "running 'rndc serve-stale on' ($n)"
+ret=0
+$RNDCCMD 10.53.0.3 serve-stale on || ret=1
+if [ $ret != 0 ]; then echo_i "failed"; fi
+status=$((status+ret))
+
+n=$((n+1))
+echo_i "check 'rndc serve-stale status' ($n)"
+ret=0
+$RNDCCMD 10.53.0.3 serve-stale status > rndc.out.test$n 2>&1 || ret=1
+grep '_default: on (rndc) (stale-answer-ttl=3 max-stale-ttl=3600 stale-refresh-time=4)' rndc.out.test$n > /dev/null || ret=1
+if [ $ret != 0 ]; then echo_i "failed"; fi
+status=$((status+ret))
+
+# Hit the fetch-limits. We burst the name server with a small batch of queries.
+# Only 2 queries are required to hit the fetch-limits. The first query will
+# start to resolve, the second one hit the fetch-limits.
 burst() {
 	num=${1}
 	rm -f burst.input.$$
 	while [ $num -gt 0 ]; do
 		num=`expr $num - 1`
-		echo "${num}.data.example A" >> burst.input.$$
+		echo "fetch${num}.example A" >> burst.input.$$
 	done
 	$PERL ../ditch.pl -p ${PORT} -s 10.53.0.3 burst.input.$$
 	rm -f burst.input.$$
 }
 
 wait_for_fetchlimits() {
-	burst 20
-	$DIG -p ${PORT} @10.53.0.3 data.example A > dig.out.test$n
+	burst 2
+	# We expect a query for nx.example to fail because fetch-limits for
+	# the domain 'example.' (and everything below) has been reached.
+	$DIG -p ${PORT} +tries=1 +timeout=1 @10.53.0.3 nx.example > dig.out.test$n
 	grep "status: SERVFAIL" dig.out.test$n > /dev/null || return 1
 }
 
@@ -2065,26 +2092,9 @@ retry_quiet 10 wait_for_fetchlimits || ret=1
 if [ $ret != 0 ]; then echo_i "failed"; fi
 status=$((status+ret))
 
-# Allow RRset to become stale.
-sleep 2
-
-# Turn on serve-stale.
-n=$((n+1))
-echo_i "running 'rndc serve-stale on' ($n)"
-ret=0
-$RNDCCMD 10.53.0.3 serve-stale on || ret=1
-if [ $ret != 0 ]; then echo_i "failed"; fi
-status=$((status+ret))
-
-n=$((n+1))
-echo_i "check 'rndc serve-stale status' ($n)"
-ret=0
-$RNDCCMD 10.53.0.3 serve-stale status > rndc.out.test$n 2>&1 || ret=1
-grep '_default: on (rndc) (stale-answer-ttl=3 max-stale-ttl=3600 stale-refresh-time=4)' rndc.out.test$n > /dev/null || ret=1
-if [ $ret != 0 ]; then echo_i "failed"; fi
-status=$((status+ret))
-
-# Expect stale data now.
+# Expect stale data now (because fetch-limits for the domain 'example.' (and
+# everything below) has been reached. But we have a stale RRset for
+# 'data.example/TXT' that can be used.
 n=$((n+1))
 ret=0
 echo_i "check stale data.example comes from cache (fetch-limits) ($n)"
-- 
2.47.3