From da45b4e7ca3a1c88c44424256c104b07710e9679 Mon Sep 17 00:00:00 2001
From: Patrick Donnelly <pdonnell@redhat.com>
Date: Thu, 21 Nov 2019 10:09:39 -0800
Subject: [PATCH] mds: release free heap pages after trim

MDS free heap space can grow to large for some workloads (like smallfile
and recursive deletes). This can cause the MDS mapped memory to grow
well beyond memory targets.

When we finally use the PriorityCache in the MDS, this will not be
necessary anymore as the PriorityCache already does this.

Fixes: https://tracker.ceph.com/issues/42938
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit af4cac5ec7bab4e5bf8936cd685f6ee8dcb38127)

Conflicts:
    src/mds/MDCache.cc
---
 src/common/options.cc |  8 +++++++-
 src/mds/MDCache.cc    | 18 +++++++++++++++---
 src/mds/MDCache.h     |  1 +
 3 files changed, 23 insertions(+), 4 deletions(-)

diff --git a/src/common/options.cc b/src/common/options.cc
index cc10924dc51..02a909e5756 100644
--- a/src/common/options.cc
+++ b/src/common/options.cc
@@ -7621,7 +7621,13 @@ std::vector<Option> get_mds_options() {
 
     Option("mds_cache_trim_interval", Option::TYPE_SECS, Option::LEVEL_ADVANCED)
     .set_default(1)
-    .set_description("interval in seconds between cache trimming"),
+    .set_description("interval in seconds between cache trimming")
+    .set_flag(Option::FLAG_RUNTIME),
+
+    Option("mds_cache_release_free_interval", Option::TYPE_SECS, Option::LEVEL_DEV)
+    .set_default(10)
+    .set_description("interval in seconds between heap releases")
+    .set_flag(Option::FLAG_RUNTIME),
 
     Option("mds_cache_size", Option::TYPE_INT, Option::LEVEL_ADVANCED)
     .set_default(0)
diff --git a/src/mds/MDCache.cc b/src/mds/MDCache.cc
index e9238391ccc..5538e70e9d6 100644
--- a/src/mds/MDCache.cc
+++ b/src/mds/MDCache.cc
@@ -163,8 +163,8 @@ MDCache::MDCache(MDSRank *m, PurgeQueue &purge_queue_) :
     while (!upkeep_trim_shutdown.load()) {
       auto now = clock::now();
       auto since = now-upkeep_last_trim;
-      auto interval = clock::duration(g_conf().get_val<std::chrono::seconds>("mds_cache_trim_interval"));
-      if (since >= interval*.90) {
+      auto trim_interval = clock::duration(g_conf().get_val<std::chrono::seconds>("mds_cache_trim_interval"));
+      if (since >= trim_interval*.90) {
         lock.unlock(); /* mds_lock -> upkeep_mutex */
         std::scoped_lock mds_lock(mds->mds_lock);
         lock.lock();
@@ -177,12 +177,24 @@ MDCache::MDCache(MDSRank *m, PurgeQueue &purge_queue_) :
           check_memory_usage();
           mds->server->recall_client_state(nullptr, Server::RecallFlags::ENFORCE_MAX);
           upkeep_last_trim = clock::now();
+          upkeep_last_trim = now = clock::now();
         } else {
           dout(10) << "cache not ready for trimming" << dendl;
         }
       } else {
-        interval -= since;
+        trim_interval -= since;
+      }
+      since = now-upkeep_last_release;
+      auto release_interval = clock::duration(g_conf().get_val<std::chrono::seconds>("mds_cache_release_free_interval"));
+      if (since >= release_interval) {
+        /* XXX not necessary once MDCache uses PriorityCache */
+        dout(10) << "releasing free memory" << dendl;
+        ceph_heap_release_free_memory();
+        upkeep_last_release = clock::now();
+      } else {
+        release_interval -= since;
       }
+      auto interval = std::min(release_interval, trim_interval);
       dout(20) << "upkeep thread waiting interval " << interval << dendl;
       upkeep_cvar.wait_for(lock, interval);
     }
diff --git a/src/mds/MDCache.h b/src/mds/MDCache.h
index 9cd83419d9e..1b37013d143 100644
--- a/src/mds/MDCache.h
+++ b/src/mds/MDCache.h
@@ -1331,6 +1331,7 @@ private:
   ceph::mutex upkeep_mutex = ceph::make_mutex("MDCache::upkeep_mutex");
   ceph::condition_variable upkeep_cvar;
   time upkeep_last_trim = time::min();
+  time upkeep_last_release = time::min();
   std::atomic<bool> upkeep_trim_shutdown{false};
 };
 
-- 
2.47.3