[sheepdog] [PATCH] sheep: remove local node from the target of SD_OP_GET_OBJ_LIST

[Liu Yuan]

On 05/17/2013 10:04 PM, MORITA Kazutaka wrote:
> Ah, I see.  The current tests can pass without getting object list from local.
> Seems that they are not enough to avoid degradation.
> 
> Anyway, I think we can get the list from local obj_list_cache instead of
> sending SD_OP_GET_OBJ_LIST.

I think either(get list from socket or direct get from cache) is okay
, it is similar to our gateway read/write. Unless we have a benefit, I'd
rather keep it as is.

Can't pass following test with your patch.

>From 87c1b4e3b2caf0bbb80522e29c0f54d68825a1dd Mon Sep 17 00:00:00 2001
From: Liu Yuan <tailai.ly at taobao.com>
Date: Fri, 17 May 2013 22:44:24 +0800
Subject: [PATCH] tests: add 061 to test list preparation of sheep recovery
 logic

Signed-off-by: Liu Yuan <tailai.ly at taobao.com>

diff --git a/tests/061 b/tests/061
new file mode 100755
index 0000000..c1488bd
--- /dev/null
+++ b/tests/061
@@ -0,0 +1,34 @@
+#!/bin/bash
+
+# Test list preparation of sheep recovery logic
+
+seq=`basename $0`
+echo "QA output created by $seq"
+
+here=`pwd`
+tmp=/tmp/$$
+status=1        # failure is the default!
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+_cleanup
+
+for i in `seq 0 2`; do
+    _start_sheep $i
+done
+
+_wait_for_sheep 3
+
+$COLLIE cluster format -c 1
+sleep 1
+
+$COLLIE vdi create test 40M -P
+
+_kill_sheep 2
+_wait_for_sheep_recovery 0
+_start_sheep 2
+_wait_for_sheep_recovery 0
+
+$COLLIE vdi check test
diff --git a/tests/061.out b/tests/061.out
new file mode 100644
index 0000000..fcc743d
--- /dev/null
+++ b/tests/061.out
@@ -0,0 +1,3 @@
+QA output created by 061
+using backend farm store
+finish check&repair test
diff --git a/tests/group b/tests/group
index 3d7f785..875f900 100644
--- a/tests/group
+++ b/tests/group
@@ -74,3 +74,4 @@
 058 auto quick cluster md
 059 auto quick vdi md
 060 auto quick cluster md
+061 auto quick cluster md