summaryrefslogtreecommitdiffstats
path: root/xlators/cluster/afr/src/afr-dir-write.c
diff options
context:
space:
mode:
authorPranith Kumar K <pkarampu@redhat.com>2012-11-27 12:34:18 +0530
committerAnand Avati <avati@redhat.com>2012-12-04 14:50:52 -0800
commit676b8793b9a77babf0275ee78eeacfa7291a8c35 (patch)
tree7f1a51c139060d43ae4e318961c8d222af760324 /xlators/cluster/afr/src/afr-dir-write.c
parentd1bb60ad794dcf0fc8bf0cf8dc4ac13349d749d4 (diff)
cluster/afr: mark new entry changelog for create/mknod failures
Problem: When create/mknod fails on some of the nodes, appropriate pending data/metadata changelogs are not assigned. This was not considered to be an issue because entry self-heal would do the assigning of appropriate changelog after creating new entries. But using the combination of rebalance and remove brick we can construct a case where a file with same name and gfid can be created in a dir with different data and link-to xattr without any changelog. Fix: When a create/mknod failure is observed mark the appropriate changelog on the new file created. Change-Id: I4c32cbf5594a13fb14deaf97ff30b2fff11cbfd6 BUG: 858212 Signed-off-by: Pranith Kumar K <pkarampu@redhat.com> Reviewed-on: http://review.gluster.org/4207 Tested-by: Gluster Build System <jenkins@build.gluster.com> Reviewed-by: Anand Avati <avati@redhat.com>
Diffstat (limited to 'xlators/cluster/afr/src/afr-dir-write.c')
-rw-r--r--xlators/cluster/afr/src/afr-dir-write.c127
1 files changed, 126 insertions, 1 deletions
diff --git a/xlators/cluster/afr/src/afr-dir-write.c b/xlators/cluster/afr/src/afr-dir-write.c
index 49cf8d4..d86b08b 100644
--- a/xlators/cluster/afr/src/afr-dir-write.c
+++ b/xlators/cluster/afr/src/afr-dir-write.c
@@ -106,11 +106,134 @@ __dir_entry_fop_common_cbk (call_frame_t *frame, int child_index,
local->fresh_children[local->success_count] = child_index;
local->success_count++;
+ local->child_errno[child_index] = 0;
+ } else {
+ local->child_errno[child_index] = op_errno;
}
local->op_errno = op_errno;
}
+int
+afr_mark_new_entry_changelog_cbk (call_frame_t *frame, void *cookie,
+ xlator_t *this,
+ int32_t op_ret, int32_t op_errno,
+ dict_t *xattr, dict_t *xdata)
+{
+ int call_count = 0;
+
+ call_count = afr_frame_return (frame);
+ if (call_count == 0) {
+ AFR_STACK_DESTROY (frame);
+ }
+ return 0;
+}
+
+void
+afr_mark_new_entry_changelog (call_frame_t *frame, xlator_t *this)
+{
+ call_frame_t *new_frame = NULL;
+ afr_local_t *local = NULL;
+ afr_local_t *new_local = NULL;
+ afr_private_t *priv = NULL;
+ dict_t **xattr = NULL;
+ int32_t **changelog = NULL;
+ int i = 0;
+ GF_UNUSED int op_errno = 0;
+
+ local = frame->local;
+ priv = this->private;
+
+ new_frame = copy_frame (frame);
+ if (!new_frame) {
+ goto out;
+ }
+
+ AFR_LOCAL_ALLOC_OR_GOTO (new_frame->local, out);
+ new_local = new_frame->local;
+ changelog = afr_matrix_create (priv->child_count, AFR_NUM_CHANGE_LOGS);
+ if (!changelog)
+ goto out;
+
+ xattr = GF_CALLOC (priv->child_count, sizeof (*xattr),
+ gf_afr_mt_dict_t);
+ if (!xattr)
+ goto out;
+ for (i = 0; i < priv->child_count; i++) {
+ if (local->child_errno[i])
+ continue;
+ xattr[i] = dict_new ();
+ if (!xattr[i])
+ goto out;
+ }
+
+ afr_prepare_new_entry_pending_matrix (changelog,
+ afr_is_errno_set,
+ local->child_errno,
+ &local->cont.dir_fop.buf,
+ priv->child_count);
+
+ new_local->pending = changelog;
+ uuid_copy (new_local->loc.gfid, local->cont.dir_fop.buf.ia_gfid);
+ new_local->loc.inode = inode_ref (local->cont.dir_fop.inode);
+ new_local->call_count = local->success_count;
+
+ for (i = 0; i < priv->child_count; i++) {
+ if (local->child_errno[i])
+ continue;
+
+ afr_set_pending_dict (priv, xattr[i], changelog, i, LOCAL_LAST);
+ STACK_WIND_COOKIE (new_frame, afr_mark_new_entry_changelog_cbk,
+ (void *) (long) i, priv->children[i],
+ priv->children[i]->fops->xattrop,
+ &new_local->loc, GF_XATTROP_ADD_ARRAY,
+ xattr[i], NULL);
+ }
+ new_frame = NULL;
+out:
+ if (new_frame)
+ AFR_STACK_DESTROY (new_frame);
+ afr_xattr_array_destroy (xattr, priv->child_count);
+ return;
+}
+
+gf_boolean_t
+afr_is_new_entry_changelog_needed (glusterfs_fop_t fop)
+{
+ glusterfs_fop_t fops[] = {GF_FOP_CREATE, GF_FOP_MKNOD, GF_FOP_NULL};
+ int i = 0;
+
+ for (i = 0; fops[i] != GF_FOP_NULL; i++) {
+ if (fop == fops[i])
+ return _gf_true;
+ }
+ return _gf_false;
+}
+
+void
+afr_dir_fop_mark_entry_pending_changelog (call_frame_t *frame, xlator_t *this)
+{
+ afr_local_t *local = NULL;
+ afr_private_t *priv = NULL;
+
+ local = frame->local;
+ priv = this->private;
+
+ if (local->op_ret < 0)
+ goto out;
+
+ if (local->success_count == priv->child_count)
+ goto out;
+
+ if (!afr_is_new_entry_changelog_needed (local->op))
+ goto out;
+
+ afr_mark_new_entry_changelog (frame, this);
+
+out:
+ local->transaction.resume (frame, this);
+}
+
void
afr_dir_fop_done (call_frame_t *frame, xlator_t *this)
{
@@ -129,7 +252,7 @@ afr_dir_fop_done (call_frame_t *frame, xlator_t *this)
local->cont.dir_fop.buf.ia_gfid);
done:
local->transaction.unwind (frame, this);
- local->transaction.resume (frame, this);
+ afr_dir_fop_mark_entry_pending_changelog (frame, this);
}
/* {{{ create */
@@ -331,6 +454,7 @@ afr_create (call_frame_t *frame, xlator_t *this,
}
UNLOCK (&priv->read_child_lock);
+ local->op = GF_FOP_CREATE;
local->cont.create.flags = flags;
local->cont.create.mode = mode;
local->cont.create.fd = fd_ref (fd);
@@ -524,6 +648,7 @@ afr_mknod (call_frame_t *frame, xlator_t *this, loc_t *loc, mode_t mode,
}
UNLOCK (&priv->read_child_lock);
+ local->op = GF_FOP_MKNOD;
local->cont.mknod.mode = mode;
local->cont.mknod.dev = dev;
local->umask = umask;