features/shard: Fix vm corruption upon fix-layout

Backport of:
> Change-Id: I8a2e97d91ba3275fbc7174a008c7234fa5295d36
> BUG: 1440051
> Reviewed on: https://review.gluster.org/17010
> (cherry-picked from commit 99c8c0b03a3368d81756440ab48091e1f2430a5f)

shard's writev implementation, as part of identifying
presence of participant shards that aren't in memory,
first sends an MKNOD on these shards, and upon EEXIST error,
looks up the shards before proceeding with the writes.

The VM corruption was caused when the following happened:
1. DHT had n subvolumes initially.
2. Upon add-brick + fix-layout, the layout of .shard changed
   although the existing shards under it were yet to be migrated
   to their new hashed subvolumes.
3. During this time, there were writes on the VM falling in regions
   of the file whose corresponding shards were already existing under
   .shard.
4. Sharding xl sent MKNOD on these shards, now creating them in their
   new hashed subvolumes although there already exist shard blocks for
   this region with valid data.
5. All subsequent writes were wound on these newly created copies.

The net outcome is that both copies of the shard didn't have the correct
data. This caused the affected VMs to be unbootable.

FIX:
For want of better alternatives in DHT, the fix changes shard fops to do
a LOOKUP before the MKNOD and upon EEXIST error, perform another lookup.

Change-Id: I8a2e97d91ba3275fbc7174a008c7234fa5295d36
BUG: 1426508
RCA'd-by: Raghavendra Gowdappa <rgowdapp@redhat.com>
Reported-by: Mahdi Adnan <mahdi.adnan@outlook.com>
Signed-off-by: Krutika Dhananjay <kdhananj@redhat.com>
Reviewed-on: https://review.gluster.org/17021
Smoke: Gluster Build System <jenkins@build.gluster.org>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.org>
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>

2 files changed, 96 insertions, 59 deletions

diff --git a/xlators/features/shard/src/shard.c b/xlators/features/shard/src/shard.c
index 6ce767bd1a6..6af02b698b2 100644
--- a/xlators/features/shard/src/shard.c
+++ b/xlators/features/shard/src/shard.c
@@ -1693,11 +1693,30 @@ shard_common_lookup_shards_cbk (call_frame_t *frame, void *cookie,
 
         if (op_ret < 0) {
                 /* Ignore absence of shards in the backend in truncate fop. */
-                if (((local->fop == GF_FOP_TRUNCATE) ||
-                    (local->fop == GF_FOP_FTRUNCATE) ||
-                    (local->fop == GF_FOP_RENAME) ||
-                    (local->fop == GF_FOP_UNLINK)) && (op_errno == ENOENT))
-                        goto done;
+                switch (local->fop) {
+                case GF_FOP_TRUNCATE:
+                case GF_FOP_FTRUNCATE:
+                case GF_FOP_RENAME:
+                case GF_FOP_UNLINK:
+                        if (op_errno == ENOENT)
+                                goto done;
+                        break;
+                case GF_FOP_WRITE:
+                case GF_FOP_READ:
+                case GF_FOP_ZEROFILL:
+                case GF_FOP_DISCARD:
+                case GF_FOP_FALLOCATE:
+                        if ((!local->first_lookup_done) &&
+                            (op_errno == ENOENT)) {
+                                local->create_count++;
+                                goto done;
+                        }
+                        break;
+                default:
+                        break;
+                }
+
+                /* else */
                 gf_msg (this->name, GF_LOG_ERROR, op_errno,
                         SHARD_MSG_LOOKUP_SHARD_FAILED, "Lookup on shard %d "
                         "failed. Base file gfid = %s", shard_block_num,
@@ -1714,6 +1733,8 @@ shard_common_lookup_shards_cbk (call_frame_t *frame, void *cookie,
 done:
         call_count = shard_call_count_return (frame);
         if (call_count == 0) {
+                if (!local->first_lookup_done)
+                        local->first_lookup_done = _gf_true;
                 if (local->op_ret < 0)
                         goto unwind;
                 else
@@ -3194,47 +3215,6 @@ next:
 }
 
 int
-shard_post_lookup_shards_readv_handler (call_frame_t *frame, xlator_t *this)
-{
-        shard_local_t *local = NULL;
-
-        local = frame->local;
-
-        if (local->op_ret < 0) {
-                SHARD_STACK_UNWIND (readv, frame, local->op_ret,
-                                    local->op_errno, NULL, 0, NULL, NULL, NULL);
-                return 0;
-        }
-
-        shard_readv_do (frame, this);
-
-        return 0;
-}
-
-int
-shard_post_mknod_readv_handler (call_frame_t *frame, xlator_t *this)
-{
-        shard_local_t *local = NULL;
-
-        local = frame->local;
-
-        if (local->op_ret < 0) {
-                SHARD_STACK_UNWIND (readv, frame, local->op_ret,
-                                    local->op_errno, NULL, 0, NULL, NULL, NULL);
-                return 0;
-        }
-
-        if (!local->eexist_count) {
-                shard_readv_do (frame, this);
-        } else {
-                local->call_count = local->eexist_count;
-                shard_common_lookup_shards (frame, this, local->loc.inode,
-                                        shard_post_lookup_shards_readv_handler);
-        }
-        return 0;
-}
-
-int
 shard_common_mknod_cbk (call_frame_t *frame, void *cookie, xlator_t *this,
                         int32_t op_ret, int32_t op_errno, inode_t *inode,
                         struct iatt *buf, struct iatt *preparent,
@@ -3264,6 +3244,7 @@ done:
         call_count = shard_call_count_return (frame);
         if (call_count == 0) {
                 SHARD_UNSET_ROOT_FS_ID (frame, local);
+                local->create_count = 0;
                 local->post_mknod_handler (frame, this);
         }
 
@@ -3394,6 +3375,55 @@ err:
 }
 
 int
+shard_post_mknod_readv_handler (call_frame_t *frame, xlator_t *this);
+
+int
+shard_post_lookup_shards_readv_handler (call_frame_t *frame, xlator_t *this)
+{
+        shard_local_t *local = NULL;
+
+        local = frame->local;
+
+        if (local->op_ret < 0) {
+                SHARD_STACK_UNWIND (readv, frame, local->op_ret,
+                                    local->op_errno, NULL, 0, NULL, NULL, NULL);
+                return 0;
+        }
+
+        if (local->create_count) {
+                shard_common_resume_mknod (frame, this,
+                                           shard_post_mknod_readv_handler);
+        } else {
+                shard_readv_do (frame, this);
+        }
+
+        return 0;
+}
+
+int
+shard_post_mknod_readv_handler (call_frame_t *frame, xlator_t *this)
+{
+        shard_local_t *local = NULL;
+
+        local = frame->local;
+
+        if (local->op_ret < 0) {
+                SHARD_STACK_UNWIND (readv, frame, local->op_ret,
+                                    local->op_errno, NULL, 0, NULL, NULL, NULL);
+                return 0;
+        }
+
+        if (!local->eexist_count) {
+                shard_readv_do (frame, this);
+        } else {
+                local->call_count = local->eexist_count;
+                shard_common_lookup_shards (frame, this, local->loc.inode,
+                                        shard_post_lookup_shards_readv_handler);
+        }
+        return 0;
+}
+
+int
 shard_post_resolve_readv_handler (call_frame_t *frame, xlator_t *this)
 {
         shard_local_t *local = NULL;
@@ -3419,9 +3449,9 @@ shard_post_resolve_readv_handler (call_frame_t *frame, xlator_t *this)
         }
 
         if (local->call_count) {
-                local->create_count = local->call_count;
-                shard_common_resume_mknod (frame, this,
-                                           shard_post_mknod_readv_handler);
+                shard_common_lookup_shards (frame, this,
+                                            local->resolver_base_inode,
+                                        shard_post_lookup_shards_readv_handler);
         } else {
                 shard_readv_do (frame, this);
         }
@@ -3573,14 +3603,11 @@ shard_readv (call_frame_t *frame, xlator_t *this, fd_t *fd, size_t size,
 
         shard_lookup_base_file (frame, this, &local->loc,
                                 shard_post_lookup_readv_handler);
-
         return 0;
-
 err:
         SHARD_STACK_UNWIND (readv, frame, -1, ENOMEM, NULL, 0, NULL, NULL,
                             NULL);
         return 0;
-
 }
 
 int
@@ -3873,6 +3900,10 @@ next:
 }
 
 int
+shard_common_inode_write_post_mknod_handler (call_frame_t *frame,
+                                             xlator_t *this);
+
+int
 shard_common_inode_write_post_lookup_shards_handler (call_frame_t *frame,
                                                      xlator_t *this)
 {
@@ -3887,7 +3918,12 @@ shard_common_inode_write_post_lookup_shards_handler (call_frame_t *frame,
                 return 0;
         }
 
-        shard_common_inode_write_do (frame, this);
+        if (local->create_count) {
+                shard_common_resume_mknod (frame, this,
+                                   shard_common_inode_write_post_mknod_handler);
+        } else {
+                shard_common_inode_write_do (frame, this);
+        }
 
         return 0;
 }
@@ -3935,11 +3971,13 @@ shard_common_inode_write_post_lookup_handler (call_frame_t *frame,
 
         local->postbuf = local->prebuf;
 
-        if (local->create_count)
-                shard_common_resume_mknod (frame, this,
-                                   shard_common_inode_write_post_mknod_handler);
-        else
+        if (local->call_count) {
+                shard_common_lookup_shards (frame, this,
+                                            local->resolver_base_inode,
+                           shard_common_inode_write_post_lookup_shards_handler);
+        } else {
                 shard_common_inode_write_do (frame, this);
+        }
 
         return 0;
 }
@@ -3959,8 +3997,6 @@ shard_common_inode_write_post_resolve_handler (call_frame_t *frame,
                 return 0;
         }
 
-        local->create_count = local->call_count;
-
         shard_lookup_base_file (frame, this, &local->loc,
                                 shard_common_inode_write_post_lookup_handler);
         return 0;
diff --git a/xlators/features/shard/src/shard.h b/xlators/features/shard/src/shard.h
index 09232a45f24..73195983aa4 100644
--- a/xlators/features/shard/src/shard.h
+++ b/xlators/features/shard/src/shard.h
@@ -255,6 +255,7 @@ typedef struct shard_local {
                 shard_lock_t *shard_lock;
         } lock;
         inode_t *resolver_base_inode;
+        gf_boolean_t first_lookup_done;
 } shard_local_t;
 
 typedef struct shard_inode_ctx {