Make this test try harder, since osd peering may interfere with a timeout.
Fixes: https://tracker.ceph.com/issues/24990
Signed-off-by: Sage Weil <sage@newdream.net>
rados_conf_set(cluster, "objecter_inject_no_watch_ping", "true");
// allow a long time here since an osd peering event will renew our
// watch.
- int left = 16 * timeout;
+ int left = 256 * timeout;
std::cout << "waiting up to " << left << " for osd to time us out ..."
<< std::endl;
while (notify_err == 0 && --left) {