From: Kefu Chai <kchai@redhat.com>
Date: Mon, 9 May 2016 07:01:46 +0000 (+0800)
Subject: osd: remove all stale osdmaps in handle_osd_map()
X-Git-Tag: v0.94.8~37^2~1
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=c2ea6dbd652782bb227b7787c486660871b2d950;p=ceph.git

osd: remove all stale osdmaps in handle_osd_map()

in a large cluster, there are better chances that the OSD fails to trim
the cached osdmap in a timely manner. and sometimes, it is just unable
to keep up with the incoming osdmap if skip_maps, so the osdmap cache
can keep building up to over 250GB in size. in this change

* publish_superblock() before trimming the osdmaps, so other osdmap
  consumers of OSDService.superblock won't access the osdmaps being
  removed.
* trim all stale osdmaps in batch of conf->osd_target_transaction_size
  if skip_maps is true. in my test, it happens when the osd only
  receives the osdmap from monitor occasionally because the osd happens
  to be chosen when monitor wants to share a new osdmap with a random
  osd.
* always use dedicated transaction(s) for trimming osdmaps. so even in
  the normal case where we are able to trim all stale osdmaps in a
  single batch, a separated transaction is used. we can piggy back
  the commits for removing maps, but we keep it this way for simplicity.
* use std::min() instead MIN() for type safety

Fixes: http://tracker.ceph.com/issues/13990
Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 369db9930887d75b498927da9c97733bff4472b6)

Conflicts:
    remove the C++11ism
---

diff --git a/src/osd/OSD.cc b/src/osd/OSD.cc
index f3e5e8ef53dc..33c42da3de95 100644
--- a/src/osd/OSD.cc
+++ b/src/osd/OSD.cc
@@ -6168,6 +6168,48 @@ void OSD::osdmap_subscribe(version_t epoch, bool force_request)
   }
 }
 
+void OSD::trim_maps(epoch_t oldest, int nreceived, bool skip_maps)
+{
+  epoch_t min = std::min(oldest, service.map_cache.cached_key_lower_bound());
+  if (superblock.oldest_map >= min)
+    return;
+
+  int num = 0;
+  ObjectStore::Transaction *t = NULL;
+  for (epoch_t e = superblock.oldest_map; e < min; ++e) {
+    dout(20) << " removing old osdmap epoch " << e << dendl;
+    if (!t) {
+      t = new ObjectStore::Transaction;
+    }
+    t->remove(META_COLL, get_osdmap_pobject_name(e));
+    t->remove(META_COLL, get_inc_osdmap_pobject_name(e));
+    superblock.oldest_map = e + 1;
+    num++;
+    if (num >= cct->_conf->osd_target_transaction_size && num >= nreceived) {
+      service.publish_superblock(superblock);
+      write_superblock(*t);
+      store->queue_transaction_and_cleanup(NULL, t);
+      t = NULL;
+      num = 0;
+      if (!skip_maps) {
+	// skip_maps leaves us with a range of old maps if we fail to remove all
+	// of them before moving superblock.oldest_map forward to the first map
+	// in the incoming MOSDMap msg. so we should continue removing them in
+	// this case, even we could do huge series of delete transactions all at
+	// once.
+	break;
+      }
+    }
+  }
+  if (num > 0) {
+    service.publish_superblock(superblock);
+    write_superblock(*t);
+    store->queue_transaction_and_cleanup(NULL, t);
+  }
+  // we should not remove the cached maps
+  assert(min <= service.map_cache.cached_key_lower_bound());
+}
+
 void OSD::handle_osd_map(MOSDMap *m)
 {
   assert(osd_lock.is_locked());
@@ -6342,20 +6384,8 @@ void OSD::handle_osd_map(MOSDMap *m)
   }
 
   if (superblock.oldest_map) {
-    int num = 0;
-    epoch_t min(
-      MIN(m->oldest_map,
-	  service.map_cache.cached_key_lower_bound()));
-    for (epoch_t e = superblock.oldest_map; e < min; ++e) {
-      dout(20) << " removing old osdmap epoch " << e << dendl;
-      t.remove(META_COLL, get_osdmap_pobject_name(e));
-      t.remove(META_COLL, get_inc_osdmap_pobject_name(e));
-      superblock.oldest_map = e+1;
-      num++;
-      if (num >= cct->_conf->osd_target_transaction_size &&
-	  (uint64_t)num > (last - first))  // make sure we at least keep pace with incoming maps
-	break;
-    }
+    // make sure we at least keep pace with incoming maps
+    trim_maps(m->oldest_map, last - first + 1, skip_maps);
   }
 
   if (!superblock.oldest_map || skip_maps)
diff --git a/src/osd/OSD.h b/src/osd/OSD.h
index 517e5e6d6ac6..4e564d3ee177 100644
--- a/src/osd/OSD.h
+++ b/src/osd/OSD.h
@@ -1659,6 +1659,7 @@ private:
 
   void wait_for_new_map(OpRequestRef op);
   void handle_osd_map(class MOSDMap *m);
+  void trim_maps(epoch_t oldest, int nreceived, bool skip_maps);
   void note_down_osd(int osd);
   void note_up_osd(int osd);