cluster/afr: Don't let NFS cache stat after writes

Problem:
Afr does post-ops after write but the stat buffer it unwinds is at the
time of write, so if nfs client caches this, it will see different
ctime when it does stat on it after post-op is done. From NFS client's
perspective it thinks the file is changed. Tar which depends on this
to be correct keeps giving 'file changed as we read it' warning.
If Afr instead has to choose to unwind after post-op, eager-lock,
delayed-post-op will have to be disabled which will lead to bad
performance for all write usecases.

Fix:
Don't let client cache stat after write.

Change-Id: Ic6062acc6e5cdd97a9c83c56bd529ec83cee8a23
BUG: 1302948
Signed-off-by: Pranith Kumar K <pkarampu@redhat.com>
Signed-off-by: Anuradha Talur <atalur@redhat.com>
Reviewed-on: http://review.gluster.org/13785
Smoke: Gluster Build System <jenkins@build.gluster.com>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Niels de Vos <ndevos@redhat.com>

1 files changed, 24 insertions, 0 deletions

diff --git a/libglusterfs/src/common-utils.c b/libglusterfs/src/common-utils.c
index 56edc9f4a2b..45886dae3ee 100644
--- a/libglusterfs/src/common-utils.c
+++ b/libglusterfs/src/common-utils.c
@@ -4425,3 +4425,27 @@ fop_enum_to_string (glusterfs_fop_t fop)
 
         return "UNKNOWNFOP";
 }
+
+gf_boolean_t
+gf_is_zero_filled_stat (struct iatt *buf)
+{
+        if (!buf)
+                return 1;
+
+        /* Do not use st_dev because it is transformed to store the xlator id
+         * in place of the device number. Do not use st_ino because by this time
+         * we've already mapped the root ino to 1 so it is not guaranteed to be
+         * 0.
+         */
+        if ((buf->ia_nlink == 0) && (buf->ia_ctime == 0))
+                return 1;
+
+        return 0;
+}
+
+void
+gf_zero_fill_stat (struct iatt *buf)
+{
+        buf->ia_nlink = 0;
+        buf->ia_ctime = 0;
+}