mon: stuck peering since warning is misleading

When osds restart or manually marked down it is common to see a HEALTH_WARN claiming that PGs have been stuck peering since awhile, even though they were active. The warning is to be issued if they really are stuck peering longer than 60s.

Fixes: https://tracker.ceph.com/issues/51688
Signed-off-by: shreyanshjain7174 <ssanchet@redhat.com>
(cherry picked from commit 48a8a6c78f5c2efa2a15bf78bed7de8394ec5370)

diff --git a/src/mon/PGMap.cc b/src/mon/PGMap.cc
index 0712e0881f2ac06e1ad41430c6f70e5fd1e9e545..326d589b705ad820ee16724e6416bb3db3044182 100644 (file)
--- a/src/mon/PGMap.cc
+++ b/src/mon/PGMap.cc
@@ -3849,32 +3849,6 @@ static void _try_mark_pg_stale(
     newstat->last_unstale = ceph_clock_now();
   }
 
-    if ((cur.state & PG_STATE_PEERING) == 0 &&
-      cur.acting_primary != -1 &&
-      osdmap.is_down(cur.acting_primary)) {
-    pg_stat_t *newstat;
-    auto q = pending_inc->pg_stat_updates.find(pgid);
-    if (q != pending_inc->pg_stat_updates.end()) {
-      if ((q->second.acting_primary == cur.acting_primary) ||
-         ((q->second.state & PG_STATE_PEERING) == 0 &&
-          q->second.acting_primary != -1 &&
-          osdmap.is_down(q->second.acting_primary))) {
-       newstat = &q->second;
-      } else {
-       // pending update is no longer down or already stale
-       return;
-      }
-    } else {
-      newstat = &pending_inc->pg_stat_updates[pgid];
-      *newstat = cur;
-    }
-    dout(10) << __func__ << " marking pg " << pgid
-            << " stale (acting_primary " << newstat->acting_primary
-            << ")" << dendl;
-    newstat->state |= PG_STATE_PEERING;
-    newstat->last_peered = ceph_clock_now();
-  }
-
 }
 
 void PGMapUpdater::check_down_pgs(