Enhance backend loading diagnostics and remount logic

- Introduced a new `loading_fail_sec` parameter in the `wait_backend_idle` function to handle prolonged loading states, improving error handling for backend readiness.
- Updated the `ensure_dlbackend_bind` function to stop SwarmUI before remounting, preventing target busy errors and ensuring consistent data mounts.
- Enhanced the `recover_errored_backends` function to account for the new remount logic, improving backend recovery processes.
- Refactored tests to validate the new loading failure conditions and ensure proper handling of backend states during diagnostics.

src/gpu_rent/session.py

@@ -178,7 +178,14 @@ def _bind_access(
         clock.mark("Idle", log)
         try:
             if tune_swarm_perf(cfg, ip, log):
-                # git reset / prior boots can drop binds; remount before restart.
+                # Stop before remount — otherwise umount fails with target busy.
+                run_ssh(
+                    cfg,
+                    ip,
+                    "sudo -n systemctl stop swarmui",
+                    check=False,
+                    timeout=120,
+                )
                 run_ssh(
                     cfg,
                     ip,
@@ -188,7 +195,7 @@ def _bind_access(
                     "mkdir -p \"$src\" \"$dst\"; "
                     "cur=$(findmnt -n -o SOURCE --target \"$dst\" 2>/dev/null || true); "
                     "if [[ \"$cur\" != \"$src\" ]]; then "
-                    "umount \"$dst\" 2>/dev/null || umount -l \"$dst\" 2>/dev/null || true; "
+                    "umount -l \"$dst\" 2>/dev/null || umount \"$dst\" 2>/dev/null || true; "
                     "mount --bind \"$src\" \"$dst\" || true; "
                     "echo remounted $dst; "
                     "fi; "