librbd/mirror: remove peer UUID of group snapshot before attempting removal

Problem:
When the number of snapshots exceeds the configured limit
(rbd_mirroring_max_mirroring_snapshots), the cleanup process remove the last
group snapshot on the primary. If this removal fails partially, the
group snapshot with some or empty image snap may still remain.

As a result:
* The snapshot is incorrectly considered for synchronization to the secondary.
* The secondary snapshot can remain forever in the incomplete state, leading
  to stuck of daemon progress.

Solution:
Before attempting to remove the group snapshot:
* Explicitly remove the peer UUID associated with the snapshot.
* This ensures that the snapshot is no longer considered eligible for
  synchronization.
* Even if the actual snapshot deletion fails partially, the snapshot is not
  tried for mirroring, avoiding daemon stalls or inconsistencies.

Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>

diff --git a/src/librbd/mirror/snapshot/GroupUnlinkPeerRequest.cc b/src/librbd/mirror/snapshot/GroupUnlinkPeerRequest.cc
index 98f14be5c528fd3a997cdf7424a0fc0643f721c7..68920f34960e073dd228b088f733db58db3d3626 100644 (file)
--- a/src/librbd/mirror/snapshot/GroupUnlinkPeerRequest.cc
+++ b/src/librbd/mirror/snapshot/GroupUnlinkPeerRequest.cc
@@ -41,8 +41,6 @@ template <typename I>
 void GroupUnlinkPeerRequest<I>::unlink_peer() {
   ldout(m_cct, 10) << "rbd_mirroring_max_mirroring_snapshots = " << m_max_snaps << dendl;
 
-  uint64_t count = 0;
-  auto unlink_snap = m_group_snaps.end();
   // First pass : cleanup snaps that have no peer_uuids or are incomplete
   for (auto peer: *m_mirror_peer_uuids){
     for (auto it = m_group_snaps.begin(); it != m_group_snaps.end(); it++) {
@@ -56,14 +54,15 @@ void GroupUnlinkPeerRequest<I>::unlink_peer() {
        (ns->mirror_peer_uuids.count(peer) != 0 &&
         ns->is_primary() &&
          it->state == cls::rbd::GROUP_SNAPSHOT_STATE_INCOMPLETE)){
-       unlink_snap = it;
-       process_snapshot(*unlink_snap, peer);
+       process_snapshot(*it, peer);
        return;
       }
     }
   }
 
   for (auto peer: *m_mirror_peer_uuids){
+    uint64_t count = 0;
+    auto unlink_snap = m_group_snaps.end();
     for (auto it = m_group_snaps.begin(); it != m_group_snaps.end(); it++) {
       auto ns = std::get_if<cls::rbd::GroupSnapshotNamespaceMirror>(
          &it->snapshot_namespace);
@@ -135,17 +134,17 @@ void GroupUnlinkPeerRequest<I>::process_snapshot(cls::rbd::GroupSnapshot group_s
                                                  std::string mirror_peer_uuid) {
   ldout(m_cct, 10) << "snap id: " << group_snap.id << dendl;
   bool found = false;
-  bool has_newer_mirror_snap = false;
 
+  m_has_newer_mirror_snap = false;
   for (auto it = m_group_snaps.begin(); it != m_group_snaps.end(); it++) {
     if (it->id  == group_snap.id) {
       found = true;
     } else if (found) {
       auto ns = std::get_if<cls::rbd::GroupSnapshotNamespaceMirror>(
-         &it->snapshot_namespace);
+          &it->snapshot_namespace);
       if (ns != nullptr) {
-       has_newer_mirror_snap = true;
-       break;
+        m_has_newer_mirror_snap = true;
+        break;
       }
     }
   }
@@ -153,11 +152,13 @@ void GroupUnlinkPeerRequest<I>::process_snapshot(cls::rbd::GroupSnapshot group_s
   if (!found) {
     ldout(m_cct, 15) << "missing snapshot: snap_id=" << group_snap.id << dendl;
     finish(-ENOENT);
-    return;  
+    return;
   }
 
-  if (has_newer_mirror_snap) {
-    remove_group_snapshot(group_snap); 
+  const auto& ns = std::get<cls::rbd::GroupSnapshotNamespaceMirror>(
+      group_snap.snapshot_namespace);
+  if (ns.mirror_peer_uuids.empty()) {
+    remove_group_snapshot(group_snap);
   } else {
     remove_peer_uuid(group_snap, mirror_peer_uuid);
   }
@@ -170,9 +171,10 @@ void GroupUnlinkPeerRequest<I>::remove_peer_uuid(
                               std::string mirror_peer_uuid) {
   ldout(m_cct, 10) << dendl;
 
-  auto aio_comp = create_rados_callback<
-    GroupUnlinkPeerRequest<I>,
-    &GroupUnlinkPeerRequest<I>::handle_remove_peer_uuid>(this);
+  auto aio_comp = create_rados_callback(
+      new LambdaContext([this, group_snap](int r) {
+        handle_remove_peer_uuid(r, group_snap);
+      }));
 
   auto ns = std::get_if<cls::rbd::GroupSnapshotNamespaceMirror>(
     &group_snap.snapshot_namespace);
@@ -188,7 +190,8 @@ void GroupUnlinkPeerRequest<I>::remove_peer_uuid(
 }
 
 template <typename I>
-void GroupUnlinkPeerRequest<I>::handle_remove_peer_uuid(int r) {
+void GroupUnlinkPeerRequest<I>::handle_remove_peer_uuid(
+    int r, cls::rbd::GroupSnapshot group_snap) {
   ldout(m_cct, 10) << "r=" << r << dendl;
 
   if (r < 0) {
@@ -197,7 +200,12 @@ void GroupUnlinkPeerRequest<I>::handle_remove_peer_uuid(int r) {
     finish(r);
     return;
   }
-  list_group_snaps();
+
+  if (m_has_newer_mirror_snap) {
+    remove_group_snapshot(group_snap);
+  } else {
+    list_group_snaps();
+  }
 }
 
 

diff --git a/src/librbd/mirror/snapshot/GroupUnlinkPeerRequest.h b/src/librbd/mirror/snapshot/GroupUnlinkPeerRequest.h
index c48d04e746bc7ac863b8f718b5dff1c4b0d217ab..c624eaf9e01a88be9c98d24a3d8b08d78b571436 100644 (file)
--- a/src/librbd/mirror/snapshot/GroupUnlinkPeerRequest.h
+++ b/src/librbd/mirror/snapshot/GroupUnlinkPeerRequest.h
@@ -55,6 +55,7 @@ private:
   Context *m_on_finish;
 
   uint64_t m_max_snaps;
+  bool m_has_newer_mirror_snap = false;
   CephContext *m_cct;
 
   std::vector<cls::rbd::GroupSnapshot> m_group_snaps;
@@ -71,7 +72,7 @@ private:
 
   void remove_peer_uuid(cls::rbd::GroupSnapshot group_snap,
                         std::string mirror_peer_uuid);
-  void handle_remove_peer_uuid(int r);
+  void handle_remove_peer_uuid(int r, cls::rbd::GroupSnapshot group_snap);
 
   void remove_group_snapshot(cls::rbd::GroupSnapshot group_snap);
   void handle_remove_group_snapshot(int r);