mon: clear pg delta after some period

If we have not pg_map updates, the delta doesn't update, and can get stuck
with the velocity right before activity stopped.  This is confusing, and
can cause incorrect health warnings about in-progress recovery.

To fix this, zero the delta if there is no activity for
'mon delta reset interval' seconds.

Fixes: #5077
Signed-off-by: Sage Weil <sage@inktank.com>

4 files changed, 18 insertions, 0 deletions

diff --git a/src/common/config_opts.h b/src/common/config_opts.h
index aa54a213fbf..776ac1c067e 100644
--- a/src/common/config_opts.h
+++ b/src/common/config_opts.h
@@ -129,6 +129,7 @@ OPTION(mon_compact_on_bootstrap, OPT_BOOL, false)  // trigger leveldb compaction
 OPTION(mon_compact_on_trim, OPT_BOOL, true)       // compact (a prefix) when we trim old states
 OPTION(mon_tick_interval, OPT_INT, 5)
 OPTION(mon_subscribe_interval, OPT_DOUBLE, 300)
+OPTION(mon_delta_reset_interval, OPT_DOUBLE, 10)   // seconds of inactivity before we reset the pg delta to 0
 OPTION(mon_osd_laggy_halflife, OPT_INT, 60*60)        // (seconds) how quickly our laggy estimations decay
 OPTION(mon_osd_laggy_weight, OPT_DOUBLE, .3)          // weight for new 'samples's in laggy estimations
 OPTION(mon_osd_adjust_heartbeat_grace, OPT_BOOL, true)    // true if we should scale based on laggy estimations
diff --git a/src/mon/PGMap.cc b/src/mon/PGMap.cc
index 36a35424a20..32ff8963a0a 100644
--- a/src/mon/PGMap.cc
+++ b/src/mon/PGMap.cc
@@ -675,6 +675,13 @@ void PGMap::recovery_summary(ostream& out) const
   }
 }
 
+void PGMap::clear_delta()
+{
+  pg_sum_delta = pool_stat_t();
+  pg_sum_deltas.clear();
+  stamp_delta = ceph_clock_now(g_ceph_context);
+}
+
 void PGMap::print_summary(ostream& out) const
 {
   std::stringstream ss;
diff --git a/src/mon/PGMap.h b/src/mon/PGMap.h
index 4794a16f030..1d0f40e8ba2 100644
--- a/src/mon/PGMap.h
+++ b/src/mon/PGMap.h
@@ -78,6 +78,8 @@ public:
   pool_stat_t pg_sum_delta;
   utime_t stamp_delta;
 
+  void clear_delta();
+
   set<pg_t> creating_pgs;   // lru: front = new additions, back = recently pinged
   map<int,set<pg_t> > creating_pgs_by_osd;
 
diff --git a/src/mon/PGMonitor.cc b/src/mon/PGMonitor.cc
index ed4833bce7a..687ef92158b 100644
--- a/src/mon/PGMonitor.cc
+++ b/src/mon/PGMonitor.cc
@@ -128,6 +128,14 @@ void PGMonitor::tick()
     }
   }
 
+  if (!pg_map.pg_sum_deltas.empty()) {
+    utime_t age = ceph_clock_now(g_ceph_context) - pg_map.stamp;
+    if (age > 2 * g_conf->mon_delta_reset_interval) {
+      dout(10) << " clearing pg_map delta (" << age << " > " << g_conf->mon_delta_reset_interval << " seconds old)" << dendl;
+      pg_map.clear_delta();
+    }
+  }
+
   dout(10) << pg_map << dendl;
 }