From: Kefu Chai <kchai@redhat.com>
Date: Thu, 11 Mar 2021 10:28:18 +0000 (+0800)
Subject: mon/OSDMonitor: restructure OSDMonitor::check_failures() loop
X-Git-Tag: v15.2.13~10^2~15^2~3
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=8b65d5aede5ebb3a1517f8ab371206e6a68385b7;p=ceph.git

mon/OSDMonitor: restructure OSDMonitor::check_failures() loop

will add a trim failures call in the loop, which mutates failure_info,
while we are still iterating this map. so have to restructure the loop
a little bit.

Fixes: https://tracker.ceph.com/issues/47380
Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 6e512b2f1e228eb808d6bff1e5c159c4d16667ef)
---

diff --git a/src/mon/OSDMonitor.cc b/src/mon/OSDMonitor.cc
index 168a9830fba7d..d776db14bec25 100644
--- a/src/mon/OSDMonitor.cc
+++ b/src/mon/OSDMonitor.cc
@@ -3099,12 +3099,13 @@ bool OSDMonitor::can_mark_in(int i)
 bool OSDMonitor::check_failures(utime_t now)
 {
   bool found_failure = false;
-  for (map<int,failure_info_t>::iterator p = failure_info.begin();
-       p != failure_info.end();
-       ++p) {
-    if (can_mark_down(p->first)) {
-      found_failure |= check_failure(now, p->first, p->second);
+  auto p = failure_info.begin();
+  while (p != failure_info.end()) {
+    auto& [target_osd, fi] = *p;
+    if (can_mark_down(target_osd)) {
+      found_failure |= check_failure(now, target_osd, fi);
     }
+    ++p;
   }
   return found_failure;
 }