Summary:
In RocksDB, keys are associated with (internal) sequence numbers which denote when the keys are written
to the database. Sequence numbers in different RocksDB instances are unrelated, thus not comparable.
It is nice if we can associate sequence numbers with their corresponding actual timestamps. One thing we can
do is to support user-defined timestamp, which allows the applications to specify the format of custom timestamps
and encode a timestamp with each key. More details can be found at https://github.com/facebook/rocksdb/wiki/User-defined-Timestamp-%28Experimental%29.
This PR provides a different but complementary approach. We can associate rocksdb snapshots (defined in
https://github.com/facebook/rocksdb/blob/7.2.fb/include/rocksdb/snapshot.h#L20) with **user-specified** timestamps.
Since a snapshot is essentially an object representing a sequence number, this PR establishes a bi-directional mapping between sequence numbers and timestamps.
In the past, snapshots are usually taken by readers. The current super-version is grabbed, and a `rocksdb::Snapshot`
object is created with the last published sequence number of the super-version. You can see that the reader actually
has no good idea of what timestamp to assign to this snapshot, because by the time the `GetSnapshot()` is called,
an arbitrarily long period of time may have already elapsed since the last write, which is when the last published
sequence number is written.
This observation motivates the creation of "timestamped" snapshots on the write path. Currently, this functionality is
exposed only to the layer of `TransactionDB`. Application can tell RocksDB to create a snapshot when a transaction
commits, effectively associating the last sequence number with a timestamp. It is also assumed that application will
ensure any two snapshots with timestamps should satisfy the following:
```
snapshot1.seq < snapshot2.seq iff. snapshot1.ts < snapshot2.ts
```
If the application can guarantee that when a reader takes a timestamped snapshot, there is no active writes going on
in the database, then we also allow the user to use a new API `TransactionDB::CreateTimestampedSnapshot()` to create
a snapshot with associated timestamp.
Code example
```cpp
// Create a timestamped snapshot when committing transaction.
txn->SetCommitTimestamp(100);
txn->SetSnapshotOnNextOperation();
txn->Commit();
// A wrapper API for convenience
Status Transaction::CommitAndTryCreateSnapshot(
std::shared_ptr<TransactionNotifier> notifier,
TxnTimestamp ts,
std::shared_ptr<const Snapshot>* ret);
// Create a timestamped snapshot if caller guarantees no concurrent writes
std::pair<Status, std::shared_ptr<const Snapshot>> snapshot = txn_db->CreateTimestampedSnapshot(100);
```
The snapshots created in this way will be managed by RocksDB with ref-counting and potentially shared with
other readers. We provide the following APIs for readers to retrieve a snapshot given a timestamp.
```cpp
// Return the timestamped snapshot correponding to given timestamp. If ts is
// kMaxTxnTimestamp, then we return the latest timestamped snapshot if present.
// Othersise, we return the snapshot whose timestamp is equal to `ts`. If no
// such snapshot exists, then we return null.
std::shared_ptr<const Snapshot> TransactionDB::GetTimestampedSnapshot(TxnTimestamp ts) const;
// Return the latest timestamped snapshot if present.
std::shared_ptr<const Snapshot> TransactionDB::GetLatestTimestampedSnapshot() const;
```
We also provide two additional APIs for stats collection and reporting purposes.
```cpp
Status TransactionDB::GetAllTimestampedSnapshots(
std::vector<std::shared_ptr<const Snapshot>>& snapshots) const;
// Return timestamped snapshots whose timestamps fall in [ts_lb, ts_ub) and store them in `snapshots`.
Status TransactionDB::GetTimestampedSnapshots(
TxnTimestamp ts_lb,
TxnTimestamp ts_ub,
std::vector<std::shared_ptr<const Snapshot>>& snapshots) const;
```
To prevent the number of timestamped snapshots from growing infinitely, we provide the following API to release
timestamped snapshots whose timestamps are older than or equal to a given threshold.
```cpp
void TransactionDB::ReleaseTimestampedSnapshotsOlderThan(TxnTimestamp ts);
```
Before shutdown, RocksDB will release all timestamped snapshots.
Comparison with user-defined timestamp and how they can be combined:
User-defined timestamp persists every key with a timestamp, while timestamped snapshots maintain a volatile
mapping between snapshots (sequence numbers) and timestamps.
Different internal keys with the same user key but different timestamps will be treated as different by compaction,
thus a newer version will not hide older versions (with smaller timestamps) unless they are eligible for garbage collection.
In contrast, taking a timestamped snapshot at a certain sequence number and timestamp prevents all the keys visible in
this snapshot from been dropped by compaction. Here, visible means (seq < snapshot and most recent).
The timestamped snapshot supports the semantics of reading at an exact point in time.
Timestamped snapshots can also be used with user-defined timestamp.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9879
Test Plan:
```
make check
TEST_TMPDIR=/dev/shm make crash_test_with_txn
```
Reviewed By: siying
Differential Revision: D35783919
Pulled By: riversand963
fbshipit-source-id: 586ad905e169189e19d3bfc0cb0177a7239d1bd4
Summary:
RocksDB snapshot already has a member unix_time_ set after
snapshot is taken. It is now exposed through GetSnapshotTime() API.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9923
Test Plan: Update unit tests
Reviewed By: riversand963
Differential Revision: D36048275
Pulled By: akankshamahajan15
fbshipit-source-id: 825210ec287deb0bc3aaa9b8e1f079f07ad686fa
Summary:
It's always annoying to find a header does not include its own
dependencies and only works when included after other includes. This
change adds `make check-headers` which validates that each header can
be included at the top of a file. Some headers are excluded e.g. because
of platform or external dependencies.
rocksdb_namespace.h had to be re-worked slightly to enable checking for
failure to include it. (ROCKSDB_NAMESPACE is a valid namespace name.)
Fixes mostly involve adding and cleaning up #includes, but for
FileTraceWriter, a constructor was out-of-lined to make a forward
declaration sufficient.
This check is not currently run with `make check` but is added to
CircleCI build-linux-unity since that one is already relatively fast.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8893
Test Plan: existing tests and resolving issues detected by new check
Reviewed By: mrambacher
Differential Revision: D30823300
Pulled By: pdillinger
fbshipit-source-id: 9fff223944994c83c105e2e6496d24845dc8e572
Summary:
When dynamically linking two binaries together, different builds of RocksDB from two sources might cause errors. To provide a tool for user to solve the problem, the RocksDB namespace is changed to a flag which can be overridden in build time.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6433
Test Plan: Build release, all and jtest. Try to build with ROCKSDB_NAMESPACE with another flag.
Differential Revision: D19977691
fbshipit-source-id: aa7f2d0972e1c31d75339ac48478f34f6cfcfb3e
Summary:
Add oldest snapshot sequence property, so we can use `db.GetProperty("rocksdb.oldest-snapshot-sequence")` to get the sequence number of the oldest snapshot.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/6228
Differential Revision: D19264145
Pulled By: maysamyabandeh
fbshipit-source-id: 67fbe5304d89cbc475bd404e30d1299f7b11c010
Summary:
Part of compaction cpu goes to processing snapshot list, the larger the list the bigger the overhead. Although the lifetime of most of the snapshots is much shorter than the lifetime of compactions, the compaction conservatively operates on the list of snapshots that it initially obtained. This patch allows the snapshot list to be updated via a callback if the compaction is taking long. This should let the compaction to continue more efficiently with much smaller snapshot list.
For simplicity, to avoid the feature is disabled in two cases: i) When more than one sub-compaction are sharing the same snapshot list, ii) when Range Delete is used in which the range delete aggregator has its own copy of snapshot list.
This fixes the reverted https://github.com/facebook/rocksdb/pull/5099 issue with range deletes.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5278
Differential Revision: D15203291
Pulled By: maysamyabandeh
fbshipit-source-id: fa645611e606aa222c7ce53176dc5bb6f259c258
Summary:
Our daily stress tests are failing after this feature. Reverting temporarily until we figure the reason for test failures.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5269
Differential Revision: D15151285
Pulled By: maysamyabandeh
fbshipit-source-id: e4002b99690a97df30d4b4b58bf0f61e9591bc6e
Summary:
Part of compaction cpu goes to processing snapshot list, the larger the list the bigger the overhead. Although the lifetime of most of the snapshots is much shorter than the lifetime of compactions, the compaction conservatively operates on the list of snapshots that it initially obtained. This patch allows the snapshot list to be updated via a callback if the compaction is taking long. This should let the compaction to continue more efficiently with much smaller snapshot list.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5099
Differential Revision: D15086710
Pulled By: maysamyabandeh
fbshipit-source-id: 7649f56c3b6b2fb334962048150142a3bf9c1a12
Summary:
WriteUnPrepared adds a virtual function, MaxUnpreparedSequenceNumber, to ReadCallback, which returns 0 unless WriteUnPrepared is enabled and the transaction has uncommitted data written to the DB. Together with snapshot sequence number, this determines the last sequence that is visible to reads.
The patch clarifies the guarantees of the GetIterator API in WriteUnPrepared transactions and make use of that to statically initialize the read callback and thus avoid the virtual call.
Furthermore it increases the minimum value for min_uncommitted from 0 to 1 as seq 0 is used only for last level keys that are committed in all snapshots.
The following benchmark shows +0.26% higher throughput in seekrandom benchmark.
Benchmark:
./db_bench --benchmarks=fillrandom --use_existing_db=0 --num=1000000 --db=/dev/shm/dbbench
./db_bench --benchmarks=seekrandom[X10] --use_existing_db=1 --db=/dev/shm/dbbench --num=1000000 --duration=60 --seek_nexts=100
seekrandom [AVG 10 runs] : 20355 ops/sec; 225.2 MB/sec
seekrandom [MEDIAN 10 runs] : 20425 ops/sec; 225.9 MB/sec
./db_bench_lessvirtual3 --benchmarks=seekrandom[X10] --use_existing_db=1 --db=/dev/shm/dbbench --num=1000000 --duration=60 --seek_nexts=100
seekrandom [AVG 10 runs] : 20409 ops/sec; 225.8 MB/sec
seekrandom [MEDIAN 10 runs] : 20487 ops/sec; 226.6 MB/sec
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5049
Differential Revision: D14366459
Pulled By: maysamyabandeh
fbshipit-source-id: ebaff8908332a5ae9af7defeadabcb624be660ef
Summary:
The vector returned by SnapshotList::GetAll could have duplicate entries if two separate snapshots have the same sequence number. However, when this vector is used in compaction the duplicate entires are of no use and could be safely ignored. Moreover not having duplicate entires simplifies reasoning in the compaction_iterator.cc code. For example when searching for the previous_snap we currently use the snapshot before the current one but the way the code uses that it expects it to be also less than the current snapshot, which would be simpler to read if there is no duplicate entry in the snapshot list.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/4860
Differential Revision: D13615502
Pulled By: maysamyabandeh
fbshipit-source-id: d45bf01213ead5f39db811f951802da6fcc3332b
Summary:
There are two issues with `VisibleToActiveSnapshot`:
1. If there are no snapshots, `oldest_snapshot` will be 0 and `VisibleToActiveSnapshot` will always return true. Since the method is used to decide whether it is safe to delete obsolete files, obsolete file won't be able to delete in this case.
2. The `auto` keyword of `auto snapshots = db_impl_->snapshots()` translate to a copy of `const SnapshotList` instead of a reference. Since copy constructor of `SnapshotList` is not defined, using the copy may yield unexpected result.
Issue 2 actually hide issue 1 from being catch by tests. During test `snapshots.empty()` can return false while it should actually be empty, and `snapshots.oldest()` return an invalid address, making `oldest_snapshot` being some random large number.
The issue was originally reported by BlobDB early adopter at Kuaishou.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/4236
Differential Revision: D9188706
Pulled By: yiwu-arbug
fbshipit-source-id: a0f2624b927cf9bf28c1bb534784fee5d106f5ea
Summary:
The is an optimization to reduce lookup in the CommitCache when querying IsInSnapshot. The optimization takes the smallest uncommitted data at the time that the snapshot was taken and if the sequence number of the read data is lower than that number it assumes the data as committed.
To implement this optimization two changes are required: i) The AddPrepared function must be called sequentially to avoid out of order insertion in the PrepareHeap (otherwise the top of the heap does not indicate the smallest prepare in future too), ii) non-2PC transactions also call AddPrepared if they do not commit in one step.
Closes https://github.com/facebook/rocksdb/pull/3649
Differential Revision: D7388630
Pulled By: maysamyabandeh
fbshipit-source-id: b79506238c17467d590763582960d4d90181c600
Summary:
…db_test
options_settable_test won't pass UBSAN so disable it.
blob_db_test fails in UBSAN as SnapshotList doesn't initialize all the fields in dummy snapshot. Fix it. I don't understand why only blob_db_test fails though.
Closes https://github.com/facebook/rocksdb/pull/3477
Differential Revision: D6928681
Pulled By: siying
fbshipit-source-id: e31dd300fcdecdfd4f6af279a0987fd0cdec5122
Summary:
Previously we store sequence number range of each blob files, and use the sequence number range to check if the file can be possibly visible by a snapshot. But it adds complexity to the code, since the sequence number is only available after a write. (The current implementation get sequence number by calling GetLatestSequenceNumber(), which is wrong.) With the patch, we are not storing sequence number range, and check if snapshot_sequence < obsolete_sequence to decide if the file is visible by a snapshot (previously we check if first_sequence <= snapshot_sequence < obsolete_sequence).
Closes https://github.com/facebook/rocksdb/pull/3274
Differential Revision: D6571497
Pulled By: yiwu-arbug
fbshipit-source-id: ca06479dc1fcd8782f6525b62b7762cd47d61909
Summary:
Blob db will keep blob file if data in the file is visible to an active snapshot. Before this patch it checks whether there is an active snapshot has sequence number greater than the earliest sequence in the file. This is problematic since we take snapshot on every read, if it keep having reads, old blob files will not be cleanup. Change to check if there is an active snapshot falls in the range of [earliest_sequence, obsolete_sequence) where obsolete sequence is
1. if data is relocated to another file by garbage collection, it is the latest sequence at the time garbage collection finish
2. otherwise, it is the latest sequence of the file
Closes https://github.com/facebook/rocksdb/pull/3087
Differential Revision: D6182519
Pulled By: yiwu-arbug
fbshipit-source-id: cdf4c35281f782eb2a9ad6a87b6727bbdff27a45
Summary:
Divide the old snapshots to two lists: a few that fit into a cached array and the rest in a vector, which is expected to be empty in normal cases. The former is to optimize concurrent reads from snapshots without requiring locks. It is done by an array of std::atomic, from which std::memory_order_acquire reads are compiled to simple read instructions in most of the x86_64 architectures.
Closes https://github.com/facebook/rocksdb/pull/2758
Differential Revision: D5660504
Pulled By: maysamyabandeh
fbshipit-source-id: 524fcf9a8e7f90a92324536456912a99aaa6740c
Summary:
Implement the main body of WritePrepared pseudo code. This includes PrepareInternal and CommitInternal, as well as AddCommitted which updates the commit map. It also provides a IsInSnapshot method that could be later called form the read path to decide if a version is in the read snapshot or it should other be skipped.
This patch lacks unit tests and does not attempt to offer an efficient implementation. The idea is that to have the API specified so that we can work on related tasks in parallel.
Closes https://github.com/facebook/rocksdb/pull/2713
Differential Revision: D5640021
Pulled By: maysamyabandeh
fbshipit-source-id: bfa7a05e8d8498811fab714ce4b9c21530514e1c
Summary:
Currently, transactions can fail even if there is no actual write conflict. This is due to relying on only the memtables to check for write-conflicts. Users have to tune memtable settings to try to avoid this, but it's hard to figure out exactly how to tune these settings.
With this diff, TransactionDB will use both memtables and SST files to determine if there are any write conflicts. This relies on the fact that BlockBasedTable stores sequence numbers for all writes that happen after any open snapshot. Also, D50295 is needed to prevent SingleDelete from disappearing writes (the TODOs in this test code will be fixed once the other diff is approved and merged).
Note that Optimistic transactions will still rely on tuning memtable settings as we do not want to read from SST while on the write thread. Also, memtable settings can still be used to reduce how often TransactionDB needs to read SST files.
Test Plan: unit tests, db bench
Reviewers: rven, yhchiang, kradhakrishnan, IslamAbdelRahman, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb, yoshinorim
Differential Revision: https://reviews.facebook.net/D50475
Summary:
D51183 was reverted due to breaking the LITE build.
This diff is the same as D51183 but with a fix for the LITE BUILD(D51693)
Test Plan: run all unit tests
Reviewers: sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D51711
Summary:
D50475 enables using SST files for transaction write-conflict checking. In order for this to work, we need to make sure not to compact out SingleDeletes when there is an earlier transaction snapshot(D50295). If there is a long-held snapshot, this could reduce the benefit of the SingleDelete optimization.
This diff allows Transactions to mark snapshots as being used for write-conflict checking. Then, during compaction, we will be able to optimize SingleDeletes better in the future.
This diff adds a flag to SnapshotImpl which is used by Transactions. This diff also passes the earliest write-conflict snapshot's sequence number to CompactionIterator. This diff does not actually change Compaction (after this diff is pushed, D50295 will be able to use this information).
Test Plan: no behavior change, ran existing tests
Reviewers: rven, kradhakrishnan, yhchiang, IslamAbdelRahman, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D51183
Summary: Implemented this simple wrapper for something else I was working on. Seemed like it makes sense to expose it instead of burying it in some random code.
Test Plan: added test
Reviewers: rven, kradhakrishnan, sdong, yhchiang
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D43293
Summary: We currently issue malloc and free inside DB mutex in GetSnapshot() and ReleaseSnapshot(). Move them out.
Test Plan:
Go through all tests
make valgrind_check
Reviewers: yhchiang, rven, IslamAbdelRahman, anthony, igor
Reviewed By: igor
Subscribers: maykov, hermanlee4, MarkCallaghan, yoshinorim, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D39753
Summary:
Couple changes:
1. instead of SnapshotList, just take a vector of snapshots
2. don't take a separate parameter is_snapshots_supported. If there are snapshots in the list, that means they are supported. I actually think we should get rid of this notion of snapshots not being supported.
3. don't pass in mutable_cf_options as a parameter. Lifetime of mutable_cf_options is a bit tricky to maintain, so it's better to not pass it in for the whole compaction job. We only really need it when we install the compaction results.
Test Plan: make check
Reviewers: sdong, rven, yhchiang
Reviewed By: yhchiang
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D36627
Summary:
When using latest clang (3.6 or 3.7/trunck) rocksdb is failing with many errors. Almost all of them are missing override errors. This diff adds missing override keyword. No manual changes.
Prerequisites: bear and clang 3.5 build with extra tools
```lang=bash
% USE_CLANG=1 bear make all # generate a compilation database http://clang.llvm.org/docs/JSONCompilationDatabase.html
% clang-modernize -p . -include . -add-override
% make format
```
Test Plan:
Make sure all tests are passing.
```lang=bash
% #Use default fb code clang.
% make check
```
Verify less error and no missing override errors.
```lang=bash
% # Have trunk clang present in path.
% ROCKSDB_NO_FBCODE=1 CC=clang CXX=clang++ make
```
Reviewers: igor, kradhakrishnan, rven, meyering, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D34077
Summary:
Add a counter in SnapshotList to show number of snapshots. Also a unix timestamp in every snapshot.
Add two DB Properties to return number of snapshots and timestamp of the oldest one.
Test Plan: Add unit test checking
Reviewers: yhchiang, rven, igor
Reviewed By: igor
Subscribers: leveldb, dhruba, MarkCallaghan
Differential Revision: https://reviews.facebook.net/D29919
1, const qualifiers on return types make no sense and will trigger a compile warning: warning: type qualifiers ignored on function return type [-Wignored-qualifiers]
2, class HistogramImpl has virtual functions and thus should have a virtual destructor
3, with some toolchain, the macro __STDC_FORMAT_MACROS is predefined and thus should be checked before define
Change-Id: I69747a03bfae88671bfbb2637c80d17600159c99
Signed-off-by: liuhuahang <liuhuahang@zerus.co>
Summary:
Change namespace from leveldb to rocksdb. This allows a single
application to link in open-source leveldb code as well as
rocksdb code into the same process.
Test Plan: compile rocksdb
Reviewers: emayanke
Reviewed By: emayanke
CC: leveldb
Differential Revision: https://reviews.facebook.net/D13287
Summary: Replace include/leveldb with include/rocksdb.
Test Plan:
make clean; make check
make clean; make release
Differential Revision: https://reviews.facebook.net/D12489
Summary:
Currently, compaction filter is run on internal key older than the oldest snapshot, which is incorrect.
Compaction filter should really be run on the most recent internal key when there is no external snapshot.
Test Plan: make check; db_stress
Reviewers: dhruba
Reviewed By: dhruba
Differential Revision: https://reviews.facebook.net/D10641
Summary:
Rocks accumulates recent writes and deletes in the in-memory memtable.
When the memtable is full, it writes the contents on the memtable to
a file in L0.
This patch removes redundant records at the time of the flush. If there
are multiple versions of the same key in the memtable, then only the
most recent one is dumped into the output file. The purging of
redundant records occur only if the most recent snapshot is earlier
than the earliest record in the memtable.
Should we switch on this feature by default or should we keep this feature
turned off in the default settings?
Test Plan: Added test case to db_test.cc
Reviewers: sheki, vamsi, emayanke, heyongqiang
Reviewed By: sheki
CC: leveldb
Differential Revision: https://reviews.facebook.net/D8991
Summary:
Specific changes:
1) Turn on -Werror so all warnings are errors
2) Fix some warnings the above now complains about
3) Add proper dependency support so changing a .h file forces a .c file
to rebuild
4) Automatically use fbcode gcc on any internal machine rather than
whatever system compiler is laying around
5) Fix jemalloc to once again be used in the builds (seemed like it
wasn't being?)
6) Fix issue where 'git' would fail in build_detect_version because of
LD_LIBRARY_PATH being set in the third-party build system
Test Plan:
make, make check, make clean, touch a header file, make sure
rebuild is expected
Reviewers: dhruba
Reviewed By: dhruba
Differential Revision: https://reviews.facebook.net/D7887
Summary:
LevelDB should delete almost-new keys when a long-open snapshot exists.
The previous behavior is to keep all versions that were created after the
oldest open snapshot. This can lead to database size bloat for
high-update workloads when there are long-open snapshots and long-open
snapshot will be used for logical backup. By "almost new" I mean that the
key was updated more than once after the oldest snapshot.
If there were two snapshots with seq numbers s1 and s2 (s1 < s2), and if
we find two instances of the same key k1 that lie entirely within s1 and
s2 (i.e. s1 < k1 < s2), then the earlier version
of k1 can be safely deleted because that version is not visible in any snapshot.
Test Plan:
unit test attached
make clean check
Differential Revision: https://reviews.facebook.net/D6999
- Replace raw slice comparison with a call to user comparator.
Added test for custom comparators.
- Fix end of namespace comments.
- Fixed bug in picking inputs for a level-0 compaction.
When finding overlapping files, the covered range may expand
as files are added to the input set. We now correctly expand
the range when this happens instead of continuing to use the
old range. For example, suppose L0 contains files with the
following ranges:
F1: a .. d
F2: c .. g
F3: f .. j
and the initial compaction target is F3. We used to search
for range f..j which yielded {F2,F3}. However we now expand
the range as soon as another file is added. In this case,
when F2 is added, we expand the range to c..j and restart the
search. That picks up file F1 as well.
This change fixes a bug related to deleted keys showing up
incorrectly after a compaction as described in Issue 44.
(Sync with upstream @25072954)
Yanqin Jin