From a1da6900ac8030dd9c156b38373837a00dbb37c0 Mon Sep 17 00:00:00 2001
From: Ravishankar N <ravishankar@redhat.com>
Date: Thu, 18 Jan 2018 14:21:57 +0530
Subject: [PATCH 153/180] afr: add quorum checks in post-op

Backport of https://review.gluster.org/#/c/18571/

afr relies on pending changelog xattrs to identify source and sinks and the
setting of these xattrs happen in post-op. So if post-op fails, we need to
unwind the write txn with a failure.

Change-Id: I0f019ac03890108324ee7672883d774918b20be1
BUG: 1384983
Signed-off-by: Ravishankar N <ravishankar@redhat.com>
Reviewed-on: https://code.engineering.redhat.com/gerrit/129219
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Tested-by: RHGS Build Bot <nigelb@redhat.com>
---
 xlators/cluster/afr/src/afr-transaction.c | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)

diff --git a/xlators/cluster/afr/src/afr-transaction.c b/xlators/cluster/afr/src/afr-transaction.c
index 7e40bba..18d2ded 100644
--- a/xlators/cluster/afr/src/afr-transaction.c
+++ b/xlators/cluster/afr/src/afr-transaction.c
@@ -534,6 +534,29 @@ afr_lock_server_count (afr_private_t *priv, afr_transaction_type type)
 /* {{{ pending */
 
 
+gf_boolean_t
+afr_post_op_has_quorum (afr_local_t *local, xlator_t *this)
+{
+        afr_private_t *priv = NULL;
+        int i = 0;
+        unsigned char *post_op_children = NULL;
+
+        priv = this->private;
+        post_op_children = alloca0 (priv->child_count);
+
+        for (i = 0; i < priv->child_count; i++) {
+                if (!local->transaction.failed_subvols[i]) {
+                        post_op_children[i] = 1;
+                }
+        }
+
+        if (afr_has_quorum (post_op_children, this)) {
+                return _gf_true;
+        }
+
+        return _gf_false;
+}
+
 int
 afr_changelog_post_op_done (call_frame_t *frame, xlator_t *this)
 {
@@ -545,6 +568,12 @@ afr_changelog_post_op_done (call_frame_t *frame, xlator_t *this)
 	priv = this->private;
         int_lock = &local->internal_lock;
 
+        /* Fail the FOP if post-op did not succeed on quorum no. of bricks. */
+        if (!afr_post_op_has_quorum (local, this)) {
+                local->op_ret = -1;
+                local->op_errno = ENOTCONN;
+        }
+
 	if (local->transaction.resume_stub) {
 		call_resume (local->transaction.resume_stub);
 		local->transaction.resume_stub = NULL;
-- 
1.8.3.1