cluster/afr: Heal directory rename without rmdir/mkdir

Problem1:
When a directory is renamed while a brick
is down entry-heal always did an rm -rf on that directory on
the sink on old location and did mkdir and created the directory
hierarchy again in the new location. This is inefficient.

Problem2:
Renamedir heal order may lead to a scenario where directory in
the new location could be created before deleting it from old
location leading to 2 directories with same gfid in posix.

Fix:
As part of heal, if oldlocation is healed first and is not present in
source-brick always rename it into a hidden directory inside the
sink-brick so that when heal is triggered in new-location shd can
rename it from this hidden directory to the new-location.

If new-location heal is triggered first and it detects that the
directory already exists in the brick, then it should skip healing the
directory until it appears in the hidden directory.

Credits: Ravi for rename-data-loss.t script

Fixes: #1211
Change-Id: I0cba2006f35cd03d314d18211ce0bd530e254843
Signed-off-by: Pranith Kumar K <pkarampu@redhat.com>

1 files changed, 40 insertions, 1 deletions

diff --git a/xlators/cluster/afr/src/afr.c b/xlators/cluster/afr/src/afr.c
index 67e0a4d10be..df7366f0a65 100644
--- a/xlators/cluster/afr/src/afr.c
+++ b/xlators/cluster/afr/src/afr.c
@@ -135,6 +135,27 @@ set_data_self_heal_algorithm(afr_private_t *priv, char *algo)
     }
 }
 
+void
+afr_handle_anon_inode_options(afr_private_t *priv, dict_t *options)
+{
+    char *volfile_id_str = NULL;
+    uuid_t anon_inode_gfid = {0};
+
+    /*If volume id is not present don't enable anything*/
+    if (dict_get_str(options, "volume-id", &volfile_id_str))
+        return;
+    GF_ASSERT(strlen(AFR_ANON_DIR_PREFIX) + strlen(volfile_id_str) <= NAME_MAX);
+    /*anon_inode_name is not supposed to change once assigned*/
+    if (!priv->anon_inode_name[0]) {
+        snprintf(priv->anon_inode_name, sizeof(priv->anon_inode_name), "%s-%s",
+                 AFR_ANON_DIR_PREFIX, volfile_id_str);
+        gf_uuid_parse(volfile_id_str, anon_inode_gfid);
+        /*Flip a bit to make sure volfile-id and anon-gfid are not same*/
+        anon_inode_gfid[0] ^= 1;
+        uuid_utoa_r(anon_inode_gfid, priv->anon_gfid_str);
+    }
+}
+
 int
 reconfigure(xlator_t *this, dict_t *options)
 {
@@ -290,6 +311,10 @@ reconfigure(xlator_t *this, dict_t *options)
         consistent_io = _gf_false;
     priv->consistent_io = consistent_io;
 
+    afr_handle_anon_inode_options(priv, options);
+
+    GF_OPTION_RECONF("use-anonymous-inode", priv->use_anon_inode, options, bool,
+                     out);
     if (priv->shd.enabled) {
         if ((priv->shd.enabled != enabled_old) ||
             (timeout_old != priv->shd.timeout))
@@ -541,7 +566,9 @@ init(xlator_t *this)
 
     GF_OPTION_INIT("consistent-metadata", priv->consistent_metadata, bool, out);
     GF_OPTION_INIT("consistent-io", priv->consistent_io, bool, out);
+    afr_handle_anon_inode_options(priv, this->options);
 
+    GF_OPTION_INIT("use-anonymous-inode", priv->use_anon_inode, bool, out);
     if (priv->quorum_count != 0)
         priv->consistent_io = _gf_false;
 
@@ -553,6 +580,9 @@ init(xlator_t *this)
         goto out;
     }
 
+    priv->anon_inode = GF_CALLOC(sizeof(unsigned char), child_count,
+                                 gf_afr_mt_char);
+
     priv->child_up = GF_CALLOC(sizeof(unsigned char), child_count,
                                gf_afr_mt_char);
 
@@ -561,7 +591,8 @@ init(xlator_t *this)
     priv->halo_child_up = GF_CALLOC(sizeof(unsigned char), child_count,
                                     gf_afr_mt_char);
 
-    if (!priv->child_up || !priv->child_latency || !priv->halo_child_up) {
+    if (!priv->child_up || !priv->child_latency || !priv->halo_child_up ||
+        !priv->anon_inode) {
         ret = -ENOMEM;
         goto out;
     }
@@ -1286,6 +1317,14 @@ struct volume_options options[] = {
      .tags = {"replicate"},
      .description = "This option exists only for backward compatibility "
                     "and configuring it doesn't have any effect"},
+    {.key = {"use-anonymous-inode"},
+     .type = GF_OPTION_TYPE_BOOL,
+     .default_value = "no",
+     .op_version = {GD_OP_VERSION_8_0},
+     .flags = OPT_FLAG_CLIENT_OPT | OPT_FLAG_SETTABLE,
+     .tags = {"replicate"},
+     .description = "Setting this option heals directory renames efficiently"},
+
     {.key = {NULL}},
 };