diff --git a/lib/wasix/src/fs/mod.rs b/lib/wasix/src/fs/mod.rs
index be813fea339..a1fe86e05d0 100644
--- a/lib/wasix/src/fs/mod.rs
+++ b/lib/wasix/src/fs/mod.rs
@@ -566,6 +566,17 @@ impl WasiFs {
         }
     }
 
+    /// We need to clear the freed FD list when the journal is replayed as it
+    /// will close lots of file descriptors which will fill the list. We clear
+    /// the list and allocate new FD's instead.
+    ///
+    /// This should only be used when the file descriptors are being managed
+    /// externally (e.g. journals)
+    pub(crate) fn clear_freed_fd_list(&self) {
+        let mut freed_fds = self.freed_fds.write().unwrap();
+        freed_fds.clear();
+    }
+
     /// Closes all the file handles.
     #[allow(clippy::await_holding_lock)]
     pub async fn close_all(&self) {
diff --git a/lib/wasix/src/syscalls/journal/restore_snapshot.rs b/lib/wasix/src/syscalls/journal/restore_snapshot.rs
index 883bd1d1812..2f2699194bc 100644
--- a/lib/wasix/src/syscalls/journal/restore_snapshot.rs
+++ b/lib/wasix/src/syscalls/journal/restore_snapshot.rs
@@ -62,6 +62,10 @@ pub unsafe fn restore_snapshot(
             .map_err(anyhow_err_to_runtime_err)?;
     }
 
+    // Reset the freed FD's so that none of them will be reused after
+    // the journal resumes
+    runner.ctx.data().state().fs.clear_freed_fd_list();
+
     // Once we get to this point we are no longer replaying the journal
     // and need to clear this flag, the reason is that restoring the
     // background threads may immediately process requests while this