libcephfs: ignore restoring the open files limit

Let's just ignore restoring the open files limit, the kernel will
defer releasing the file descriptors and then the process will be
possibly reachthe open files limit.

Fixes: https://tracker.ceph.com/issues/43039
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit c871d68)

diff --git a/src/test/libcephfs/test.cc b/src/test/libcephfs/test.cc
index 857d5440818101b896e7d3ad7a6e16fd70a30951..f76c871c9a3c7a0d7dff29864c87354bfb176f42 100644 (file)
--- a/src/test/libcephfs/test.cc
+++ b/src/test/libcephfs/test.cc
@@ -2083,7 +2083,13 @@ TEST(LibCephFS, ShutdownRace)
 
   for (int i = 0; i < nthreads; ++i)
     threads[i].join();
-  ASSERT_EQ(setrlimit(RLIMIT_NOFILE, &rold), 0);
+  /*
+   * Let's just ignore restoring the open files limit,
+   * the kernel will defer releasing the file descriptors
+   * and then the process will be possibly reachthe open
+   * files limit. More detail, please see tracer#43039
+   */
+//  ASSERT_EQ(setrlimit(RLIMIT_NOFILE, &rold), 0);
 }
 
 static void get_current_time_utimbuf(struct utimbuf *utb)