From: shreyanshjain7174 <ssanchet@redhat.com>
Date: Mon, 26 Dec 2022 04:43:34 +0000 (-0600)
Subject: mon: stuck peering since warning is misleading
X-Git-Tag: v19.0.0~848^2
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=48a8a6c78f5c2efa2a15bf78bed7de8394ec5370;p=ceph.git

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>
---

diff --git a/src/mon/PGMap.cc b/src/mon/PGMap.cc
index 33613dd4a6a4..33268180f3da 100644
--- a/src/mon/PGMap.cc
+++ b/src/mon/PGMap.cc
@@ -3857,32 +3857,6 @@ 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(