From 4960f579a234f9984d73767fde073c419e884c17 Mon Sep 17 00:00:00 2001
From: Neha <nojha@redhat.com>
Date: Wed, 11 Dec 2019 23:47:19 +0000
Subject: [PATCH] osd/PeeringState.cc: skip peer_purged when discovering all
 missing

We hit a couple of bugs because in discover_all_missing() we send
pg_query to an OSD that was marked stray and already got purged. This results
in a state machine crash on the purged OSD. Fix this by skipping any
purged peers.

Fixes: https://tracker.ceph.com/issues/41317
Fixes: https://tracker.ceph.com/issues/40963
Signed-off-by: Neha Ojha <nojha@redhat.com>
---
 src/osd/PeeringState.cc | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/src/osd/PeeringState.cc b/src/osd/PeeringState.cc
index 99534bcc620..2d3450646cb 100644
--- a/src/osd/PeeringState.cc
+++ b/src/osd/PeeringState.cc
@@ -2225,6 +2225,11 @@ bool PeeringState::discover_all_missing(
       continue;
     }
 
+    if (peer_purged.count(peer)) {
+      psdout(20) << __func__ << " skipping purged osd." << peer << dendl;
+      continue;
+    }
+
     map<pg_shard_t, pg_info_t>::const_iterator iter = peer_info.find(peer);
     if (iter != peer_info.end() &&
         (iter->second.is_empty() || iter->second.dne())) {
-- 
2.39.5