From: xie xingguo <xie.xingguo@zte.com.cn>
Date: Mon, 23 Apr 2018 00:50:21 +0000 (+0800)
Subject: osd/PG: perfer EC async_recovery_targets in reverse order of cost
X-Git-Tag: v13.1.0~151^2
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=refs%2Fpull%2F21588%2Fhead;p=ceph.git

osd/PG: perfer EC async_recovery_targets in reverse order of cost

This is a follow-up fix of https://github.com/ceph/ceph/pull/21578,
in which I forget that erasure-coded-pools share the same logic
when determining the async_recovery_targets..

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
---

diff --git a/src/osd/PG.cc b/src/osd/PG.cc
index 3d6abdff24ed..32d83388e426 100644
--- a/src/osd/PG.cc
+++ b/src/osd/PG.cc
@@ -1528,8 +1528,9 @@ void PG::choose_async_recovery_ec(const map<pg_shard_t, pg_info_t> &all_info,
            << dendl;
 
   // take out as many osds as we can for async recovery, in order of cost
-  for (auto weighted_shard : candidates_by_cost) {
-    pg_shard_t cur_shard = weighted_shard.second;
+  for (auto rit = candidates_by_cost.rbegin();
+       rit != candidates_by_cost.rend(); ++rit) {
+    pg_shard_t cur_shard = rit->second;
     vector<int> candidate_want(*want);
     candidate_want[cur_shard.shard.id] = CRUSH_ITEM_NONE;
     if (recoverable_and_ge_min_size(candidate_want)) {