qa: wait rank 0 to become up:active state before mounting fuse client

When setting the ec pool to the layout the filesystem may not be
ready, so when mounting a fuse client it will fail. To fix this we
need to wait at least the rank 0 to be in up:active state.

Fixes: https://tracker.ceph.com/issues/55824
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 72194627c19bcb97b0a70d28e63185c9e9c015aa)

diff --git a/qa/tasks/cephfs/filesystem.py b/qa/tasks/cephfs/filesystem.py
index 63b02cfcd9e2175d8cab60325f9a2c9dfd0adf8f..19343b07cf9d1ffbae9408be4695b52372d6cc3c 100644 (file)
--- a/qa/tasks/cephfs/filesystem.py
+++ b/qa/tasks/cephfs/filesystem.py
@@ -734,6 +734,11 @@ class Filesystem(MDSCluster):
     def run_client_payload(self, cmd):
         # avoid circular dep by importing here:
         from tasks.cephfs.fuse_mount import FuseMount
+
+        # Wait for at MDS daemons to be ready before mounting the
+        # ceph-fuse client in run_client_payload()
+        self.wait_for_daemons()
+
         d = misc.get_testdir(self._ctx)
         m = FuseMount(self._ctx, {}, d, "admin", self.client_remote, cephfs_name=self.name)
         m.mount_wait()