From: Sage Weil <sage@redhat.com>
Date: Wed, 31 Jul 2019 05:19:39 +0000 (-0500)
Subject: mon/HealthMonitor: automatically clear non-sticky mutes when alert clears
X-Git-Tag: v15.1.0~1877^2~18
X-Git-Url: http://git.apps.os.sepia.ceph.com/?a=commitdiff_plain;h=70db3f025a02b599522a1b78157ae99d7208f889;p=ceph-ci.git

mon/HealthMonitor: automatically clear non-sticky mutes when alert clears

If the alert goes away, drop the mute.

Signed-off-by: Sage Weil <sage@redhat.com>
---

diff --git a/src/mon/HealthMonitor.cc b/src/mon/HealthMonitor.cc
index b028adb8afb..3f8ed8f9ab5 100644
--- a/src/mon/HealthMonitor.cc
+++ b/src/mon/HealthMonitor.cc
@@ -364,6 +364,8 @@ bool HealthMonitor::check_mutes()
 {
   bool changed = true;
   auto now = ceph_clock_now();
+  health_check_map_t all;
+  gather_all_health_checks(&all);
   auto p = pending_mutes.begin();
   while (p != pending_mutes.end()) {
     if (p->second.ttl != utime_t() &&
@@ -372,6 +374,12 @@ bool HealthMonitor::check_mutes()
 			<< " cleared (passed TTL " << p->second.ttl << ")";
       p = pending_mutes.erase(p);
       changed = true;
+    } else if (!p->second.sticky &&
+	       all.checks.count(p->first) == 0) {
+      mon->clog->info() << "Health alert mute " << p->first
+			<< " cleared (health alert cleared)";
+      p = pending_mutes.erase(p);
+      changed = true;
     } else {
       ++p;
     }