update sources to v12.1.0

[ceph.git] / ceph / src / osd / PGBackend.cc

diff --git a/ceph/src/osd/PGBackend.cc b/ceph/src/osd/PGBackend.cc
index fab9ee826642c676747668e9f08f706a8fd38d65..97a7f07ba326ea5cb6dfd1355b737ba6655cfe5c 100644 (file)
--- a/ceph/src/osd/PGBackend.cc
+++ b/ceph/src/osd/PGBackend.cc
@@ -620,10 +620,21 @@ map<pg_shard_t, ScrubMap *>::const_iterator
   eversion_t auth_version;
   bufferlist auth_bl;
 
-  map<pg_shard_t, ScrubMap *>::const_iterator auth = maps.end();
+  // Create list of shards with primary last so it will be auth copy all
+  // other things being equal.
+  list<pg_shard_t> shards;
   for (map<pg_shard_t, ScrubMap *>::const_iterator j = maps.begin();
        j != maps.end();
        ++j) {
+    if (j->first == get_parent()->whoami_shard())
+      continue;
+    shards.push_back(j->first);
+  }
+  shards.push_back(get_parent()->whoami_shard());
+
+  map<pg_shard_t, ScrubMap *>::const_iterator auth = maps.end();
+  for (auto &l : shards) {
+    map<pg_shard_t, ScrubMap *>::const_iterator j = maps.find(l);
     map<hobject_t, ScrubMap::object>::iterator i =
       j->second->objects.find(obj);
     if (i == j->second->objects.end()) {
@@ -647,6 +658,8 @@ map<pg_shard_t, ScrubMap *>::const_iterator
     object_info_t oi;
     bufferlist bl;
     map<string, bufferptr>::iterator k;
+    SnapSet ss;
+    bufferlist ss_bl;
 
     if (i->second.stat_error) {
       shard_info.set_stat_error();
@@ -674,6 +687,12 @@ map<pg_shard_t, ScrubMap *>::const_iterator
       goto out;
     }
 
+    if (oi.soid != obj) {
+      shard_info.set_oi_attr_corrupted();
+      error_string += " oi_attr_corrupted";
+      goto out;
+    }
+
     if (auth_version != eversion_t()) {
       if (!object_error.has_object_info_inconsistency() && !(bl == auth_bl)) {
        object_error.set_object_info_inconsistency();
@@ -686,6 +705,23 @@ map<pg_shard_t, ScrubMap *>::const_iterator
     if (i->second.read_error || i->second.ec_hash_mismatch || i->second.ec_size_mismatch)
       goto out;
 
+    // We don't set errors here for snapset, but we won't pick an auth copy if the
+    // snapset is missing or won't decode.
+    if (obj.is_head() || obj.is_snapdir()) {
+      k = i->second.attrs.find(SS_ATTR);
+      if (k == i->second.attrs.end()) {
+       goto out;
+      }
+      ss_bl.push_back(k->second);
+      try {
+       bufferlist::iterator bliter = ss_bl.begin();
+       ::decode(ss, bliter);
+      } catch (...) {
+       // invalid snapset, probably corrupt
+       goto out;
+      }
+    }
+
     if (auth_version == eversion_t() || oi.version > auth_version ||
         (oi.version == auth_version && dcount(oi) > dcount(*auth_oi))) {
       auth = j;
@@ -750,6 +786,7 @@ void PGBackend::be_compare_scrubmaps(
       be_select_auth_object(*k, maps, &auth_oi, shard_map, object_error);
 
     list<pg_shard_t> auth_list;
+    set<pg_shard_t> object_errors;
     if (auth == maps.end()) {
       object_error.set_version(0);
       object_error.set_auth_missing(*k, maps, shard_map, shallow_errors, deep_errors);
@@ -793,6 +830,10 @@ void PGBackend::be_compare_scrubmaps(
          if (found)
            errorstream << pgid << " shard " << j->first << ": soid " << *k
                      << " " << ss.str() << "\n";
+       } else if (found) {
+         // Track possible shard to use as authoritative, if needed
+         // There are errors, without identifying the shard
+         object_errors.insert(j->first);
        } else {
          // XXX: The auth shard might get here that we don't know
          // that it has the "correct" data.
@@ -810,10 +851,25 @@ void PGBackend::be_compare_scrubmaps(
     }
 
     if (auth_list.empty()) {
-      errorstream << pgid.pgid << " soid " << *k
+      if (object_errors.empty()) {
+        errorstream << pgid.pgid << " soid " << *k
                  << ": failed to pick suitable auth object\n";
-      goto out;
+        goto out;
+      }
+      // Object errors exist and nothing in auth_list
+      // Prefer the auth shard otherwise take first from list.
+      pg_shard_t shard;
+      if (object_errors.count(auth->first)) {
+       shard = auth->first;
+      } else {
+       shard = *(object_errors.begin());
+      }
+      auth_list.push_back(shard);
+      object_errors.erase(shard);
     }
+    // At this point auth_list is populated, so we add the object errors shards
+    // as inconsistent.
+    cur_inconsistent.insert(object_errors.begin(), object_errors.end());
     if (!cur_missing.empty()) {
       missing[*k] = cur_missing;
     }