librbd: retry lock requests periodically until acquired

If the exclusive lock owner acks the lock release request but crashes
before it actually releases the lock, the requestor will wait forever.
Therefore, after a certain timeout, retry the request again until it
succeeds.

Fixes: #11537
Backport: hammer
Signed-off-by: Jason Dillaman <dillaman@redhat.com>

diff --git a/src/librbd/ImageWatcher.cc b/src/librbd/ImageWatcher.cc
index 82e45db1693781dab3b37e9b9aab7dc91d8a876b..6cfa6d7c5626ca675652d98b1e280b3174679bfe 100644 (file)
--- a/src/librbd/ImageWatcher.cc
+++ b/src/librbd/ImageWatcher.cc
@@ -626,6 +626,13 @@ void ImageWatcher::notify_request_lock() {
     lderr(m_image_ctx.cct) << "error requesting lock: " << cpp_strerror(r)
                           << dendl;
     schedule_retry_aio_requests(true);
+  } else {
+    // lock owner acked -- but resend if we don't see them release the lock
+    int retry_timeout = m_image_ctx.cct->_conf->client_notify_timeout;
+    FunctionContext *ctx = new FunctionContext(
+      boost::bind(&ImageWatcher::notify_request_lock, this));
+    m_task_finisher->add_event_after(TASK_CODE_REQUEST_LOCK,
+                                     retry_timeout, ctx);
   }
 }
 
@@ -777,6 +784,7 @@ void ImageWatcher::handle_payload(const HeaderUpdatePayload &payload,
 void ImageWatcher::handle_payload(const AcquiredLockPayload &payload,
                                   bufferlist *out) {
   ldout(m_image_ctx.cct, 10) << "image exclusively locked announcement" << dendl;
+  m_task_finisher->cancel(TASK_CODE_REQUEST_LOCK);
   if (payload.client_id.is_valid()) {
     Mutex::Locker l(m_owner_client_id_lock);
     if (payload.client_id == m_owner_client_id) {
@@ -796,6 +804,7 @@ void ImageWatcher::handle_payload(const AcquiredLockPayload &payload,
 void ImageWatcher::handle_payload(const ReleasedLockPayload &payload,
                                   bufferlist *out) {
   ldout(m_image_ctx.cct, 10) << "exclusive lock released" << dendl;
+  m_task_finisher->cancel(TASK_CODE_REQUEST_LOCK);
   if (payload.client_id.is_valid()) {
     Mutex::Locker l(m_owner_client_id_lock);
     if (payload.client_id != m_owner_client_id) {