SERVER-30378 Add DeferredWriter.

1 files changed, 173 insertions, 0 deletions

diff --git a/src/mongo/db/concurrency/deferred_writer.cpp b/src/mongo/db/concurrency/deferred_writer.cpp
new file mode 100644
index 00000000000..0ffa4a7c353
--- /dev/null
+++ b/src/mongo/db/concurrency/deferred_writer.cpp
@@ -0,0 +1,173 @@
+/**
+ *    Copyright (C) 2017 MongoDB Inc.
+ *
+ *    This program is free software: you can redistribute it and/or  modify
+ *    it under the terms of the GNU Affero General Public License, version 3,
+ *    as published by the Free Software Foundation.
+ *
+ *    This program is distributed in the hope that it will be useful,
+ *    but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *    GNU Affero General Public License for more details.
+ *
+ *    You should have received a copy of the GNU Affero General Public License
+ *    along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ *    As a special exception, the copyright holders give permission to link the
+ *    code of portions of this program with the OpenSSL library under certain
+ *    conditions as described in each individual source file and distribute
+ *    linked combinations including the program with the OpenSSL library. You
+ *    must comply with the GNU Affero General Public License in all respects
+ *    for all of the code used other than as permitted herein. If you modify
+ *    file(s) with this exception, you may extend this exception to your
+ *    version of the file(s), but you are not obligated to do so. If you do not
+ *    wish to do so, delete this exception statement from your version. If you
+ *    delete this exception statement from all source files in the program,
+ *    then also delete it in the license file.
+ */
+
+#define MONGO_LOG_DEFAULT_COMPONENT ::mongo::logger::LogComponent::kWrite
+
+#include "mongo/db/concurrency/deferred_writer.h"
+#include "mongo/db/catalog/create_collection.h"
+#include "mongo/db/client.h"
+#include "mongo/db/concurrency/write_conflict_exception.h"
+#include "mongo/db/db_raii.h"
+#include "mongo/db/operation_context.h"
+#include "mongo/util/concurrency/idle_thread_block.h"
+#include "mongo/util/concurrency/thread_pool.h"
+#include "mongo/util/log.h"
+
+namespace mongo {
+
+namespace {
+auto kLogInterval = stdx::chrono::hours(1);
+}
+
+void DeferredWriter::_logFailure(const Status& status) {
+    if (TimePoint::clock::now() - _lastLogged > kLogInterval) {
+        log() << "Unable to write to collection " << _nss.toString() << ": " << status.toString();
+        _lastLogged = stdx::chrono::system_clock::now();
+    }
+}
+
+Status DeferredWriter::_makeCollection(OperationContext* opCtx) {
+    BSONObjBuilder builder;
+    builder.append("create", _nss.coll());
+    builder.appendElements(_collectionOptions.toBSON());
+    try {
+        return createCollection(opCtx, _nss.db().toString(), builder.obj().getOwned());
+    } catch (const DBException& exception) {
+        return exception.toStatus();
+    }
+}
+
+StatusWith<std::unique_ptr<AutoGetCollection>> DeferredWriter::_getCollection(
+    OperationContext* opCtx) {
+    std::unique_ptr<AutoGetCollection> agc;
+    agc = stdx::make_unique<AutoGetCollection>(opCtx, _nss, MODE_IX);
+
+    while (!agc->getCollection()) {
+        // Release the previous AGC's lock before trying to rebuild the collection.
+        agc.reset();
+        Status status = _makeCollection(opCtx);
+
+        if (!status.isOK()) {
+            return status;
+        }
+
+        agc = stdx::make_unique<AutoGetCollection>(opCtx, _nss, MODE_IX);
+    }
+
+    return std::move(agc);
+}
+
+void DeferredWriter::_worker(InsertStatement stmt) {
+    auto uniqueOpCtx = Client::getCurrent()->makeOperationContext();
+    OperationContext* opCtx = uniqueOpCtx.get();
+    auto result = _getCollection(opCtx);
+
+    if (!result.isOK()) {
+        _logFailure(result.getStatus());
+        return;
+    }
+
+    auto agc = std::move(result.getValue());
+
+    Collection& collection = *agc->getCollection();
+
+    Status status = writeConflictRetry(opCtx, "deferred insert", _nss.ns(), [&] {
+        WriteUnitOfWork wuow(opCtx);
+        Status status = collection.insertDocument(opCtx, stmt, nullptr, false);
+        if (!status.isOK()) {
+            return status;
+        }
+
+        wuow.commit();
+        return Status::OK();
+    });
+
+    stdx::lock_guard<stdx::mutex> lock(_mutex);
+
+    _numBytes -= stmt.doc.objsize();
+
+    // If a write to a deferred collection fails, periodically tell the log.
+    if (!status.isOK()) {
+        _logFailure(status);
+    }
+}
+
+DeferredWriter::DeferredWriter(NamespaceString nss, CollectionOptions opts, int64_t maxSize)
+    : _collectionOptions(opts),
+      _maxNumBytes(maxSize),
+      _nss(nss),
+      _numBytes(0),
+      _lastLogged(TimePoint::clock::now() - kLogInterval) {}
+
+DeferredWriter::~DeferredWriter() {}
+
+void DeferredWriter::startup(std::string workerName) {
+    // We should only start up once.
+    invariant(!_pool);
+    ThreadPool::Options options;
+    options.poolName = "deferred writer pool";
+    options.threadNamePrefix = workerName;
+    options.minThreads = 0;
+    options.maxThreads = 1;
+    options.onCreateThread = [](const std::string& name) { Client::initThread(name); };
+    _pool = stdx::make_unique<ThreadPool>(options);
+    _pool->startup();
+}
+
+void DeferredWriter::shutdown(void) {
+    // If we never allocated the pool, no cleanup is necessary.
+    if (!_pool) {
+        return;
+    }
+
+    _pool->waitForIdle();
+    _pool->shutdown();
+    _pool->join();
+}
+
+bool DeferredWriter::insertDocument(BSONObj obj) {
+    // We can't insert documents if we haven't been started up.
+    invariant(_pool);
+
+    stdx::lock_guard<stdx::mutex> lock(_mutex);
+
+    // Check if we're allowed to insert this object.
+    if (_numBytes + obj.objsize() >= _maxNumBytes) {
+        // If not, drop it.  We always drop new entries rather than old ones; that way the caller
+        // knows at the time of the call that the entry was dropped.
+        return false;
+    }
+
+    // Add the object to the buffer.
+    _numBytes += obj.objsize();
+    fassertStatusOK(40588,
+                    _pool->schedule([this, obj] { _worker(InsertStatement(obj.getOwned())); }));
+    return true;
+}
+
+}  // namespace mongo