From: Kefu Chai <kchai@redhat.com>
Date: Mon, 29 May 2017 16:40:25 +0000 (+0800)
Subject: mgr: reset pending_inc after applying it
X-Git-Tag: ses5-milestone6~8^2~19^2~19
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=aad1afcb96de1483ecaf93adf473d5732246ea7a;p=ceph.git

mgr: reset pending_inc after applying it

we cannot apply pending_inc twice and expect the result is the same. in
other words, pg_map.apply_incremental(pending_inc) is not an idempotent
operation.

Signed-off-by: Kefu Chai <kchai@redhat.com>
---

diff --git a/src/mgr/ClusterState.cc b/src/mgr/ClusterState.cc
index 3ca97499cd9a..60612cb90b50 100644
--- a/src/mgr/ClusterState.cc
+++ b/src/mgr/ClusterState.cc
@@ -96,6 +96,7 @@ void ClusterState::update_delta_stats()
   pending_inc.version = pg_map.version + 1; // to make apply_incremental happy
   dout(10) << " v" << pending_inc.version << dendl;
   pg_map.apply_incremental(g_ceph_context, pending_inc);
+  pending_inc = PGMap::Incremental();
 }
 
 void ClusterState::notify_osdmap(const OSDMap &osd_map)
@@ -122,7 +123,7 @@ void ClusterState::notify_osdmap(const OSDMap &osd_map)
 			       need_check_down_pg_osds, &pending_inc);
 
   pg_map.apply_incremental(g_ceph_context, pending_inc);
-
+  pending_inc = PGMap::Incremental();
   // TODO: Complete the separation of PG state handling so
   // that a cut-down set of functionality remains in PGMonitor
   // while the full-blown PGMap lives only here.