From: Alexey Lyashkov <alexey.lyashkov@hpe.com>
Date: Thu, 30 Dec 2021 13:04:59 +0000 (+0300)
Subject: LU-15403 tests: fix some false alarm
X-Git-Tag: 2.15.0-RC1~67
X-Git-Url: https://git.whamcloud.com/gitweb?a=commitdiff_plain;h=ed31c0cf4b09f4b877f79970de258008307e59bc;p=fs%2Flustre-release.git

LU-15403 tests: fix some false alarm

Current test implementation have a two bugs.
1) client mgc llogs processing is async to mount,
so we can start a lock check before all locks processed.
it caused a false alarm. Same story with several client mounts.

2) Server locks counting is unsafe, as it include an other server
locks. so any servers reconnect may cause a false alarm.

Let's fix it.

HPe-bug-id: LUS-8326
Test-Parameters: trivial testlist=sanityn
Signed-off-by: Alexey Lyashkov <alexey.lyashkov@hpe.com>
Change-Id: I59d6e5deb79ca9f040385231738b8698a3309e8e
Reviewed-on: https://review.whamcloud.com/45958
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
---

diff --git a/lustre/tests/sanity.sh b/lustre/tests/sanity.sh
index 64cbe5c..6d96f85 100755
--- a/lustre/tests/sanity.sh
+++ b/lustre/tests/sanity.sh
@@ -27970,6 +27970,8 @@ run_test 900 "umount should not race with any mgc requeue thread"
 
 # LUS-6253/LU-11185
 test_901() {
+	local old
+	local count
 	local oldc
 	local newc
 	local olds
@@ -27978,16 +27980,28 @@ test_901() {
 
 	# some get_param have a bug to handle dot in param name
 	cancel_lru_locks MGC
-	oldc=$($LCTL get_param -n 'ldlm.namespaces.MGC*.lock_count')
-	olds=$(do_facet mgs $LCTL get_param -n 'ldlm.namespaces.MGS*.lock_count')
+	old=$(mount -t lustre | wc -l)
+	# 1 config+sptlrpc
+	# 2 params
+	# 3 nodemap
+	# 4 IR
+	old=$((old * 4))
+	oldc=0
+	count=0
+	while [ $old -ne $oldc ]; do
+		oldc=$($LCTL get_param -n 'ldlm.namespaces.MGC*.lock_count')
+		sleep 1
+		((count++))
+		if [ $count -ge $TIMEOUT ]; then
+			error "too large timeout"
+		fi
+	done
 	umount_client $MOUNT || error "umount failed"
 	mount_client $MOUNT || error "mount failed"
 	cancel_lru_locks MGC
 	newc=$($LCTL get_param -n 'ldlm.namespaces.MGC*.lock_count')
-	news=$(do_facet mgs $LCTL get_param -n 'ldlm.namespaces.MGS*.lock_count')
 
 	[ $oldc -lt $newc ] && error "mgc lock leak ($oldc != $newc)"
-	[ $olds -lt $news ] && error "mgs lock leak ($olds != $news)"
 
 	return 0
 }