Summary:
There are a number of fixes in this PR (with most bugs found via the added stress tests):
1. Re-enable reseek optimization. This was initially disabled to avoid infinite loops in https://github.com/facebook/rocksdb/pull/3955 but this can be resolved by remembering not to reseek after a reseek has already been done. This problem only affects forward iteration in `DBIter::FindNextUserEntryInternal`, as we already disable reseeking in `DBIter::FindValueForCurrentKeyUsingSeek`.
2. Verify that ReadOption.snapshot can be safely used for iterator creation. Some snapshots would not give correct results because snaphsot validation would not be enforced, breaking some assumptions in Prev() iteration.
3. In the non-snapshot Get() case, reads done at `LastPublishedSequence` may not be enough, because unprepared sequence numbers are not published. Use `std::max(published_seq, max_visible_seq)` to do lookups instead.
4. Add stress test to test reading own writes.
5. Minor bug in the allow_concurrent_memtable_write case where we forgot to pass in batch_per_txn_.
6. Minor performance optimization in `CalcMaxUnpreparedSequenceNumber` by assigning by reference instead of value.
7. Add some more comments everywhere.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5573
Differential Revision: D16276089
Pulled By: lth
fbshipit-source-id: 18029c944eb427a90a87dee76ac1b23f37ec1ccb
Summary:
Added log_readahead_size option to control prefetching for Log::Reader.
This is mostly useful for reading a remotely located log, as it can save the number of round-trips when reading it.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5592
Differential Revision: D16362989
Pulled By: elipoz
fbshipit-source-id: c5d4d5245a44008cd59879640efff70c091ad3e8
Summary:
Refresh of the earlier change here - https://github.com/facebook/rocksdb/issues/5135
This is a review request for code change needed for - https://github.com/facebook/rocksdb/issues/3469
"Add support for taking snapshot of a column family and creating column family from a given CF snapshot"
We have an implementation for this that we have been testing internally. We have two new APIs that together provide this functionality.
(1) ExportColumnFamily() - This API is modelled after CreateCheckpoint() as below.
// Exports all live SST files of a specified Column Family onto export_dir,
// returning SST files information in metadata.
// - SST files will be created as hard links when the directory specified
// is in the same partition as the db directory, copied otherwise.
// - export_dir should not already exist and will be created by this API.
// - Always triggers a flush.
virtual Status ExportColumnFamily(ColumnFamilyHandle* handle,
const std::string& export_dir,
ExportImportFilesMetaData** metadata);
Internally, the API will DisableFileDeletions(), GetColumnFamilyMetaData(), Parse through
metadata, creating links/copies of all the sst files, EnableFileDeletions() and complete the call by
returning the list of file metadata.
(2) CreateColumnFamilyWithImport() - This API is modeled after IngestExternalFile(), but invoked only during a CF creation as below.
// CreateColumnFamilyWithImport() will create a new column family with
// column_family_name and import external SST files specified in metadata into
// this column family.
// (1) External SST files can be created using SstFileWriter.
// (2) External SST files can be exported from a particular column family in
// an existing DB.
// Option in import_options specifies whether the external files are copied or
// moved (default is copy). When option specifies copy, managing files at
// external_file_path is caller's responsibility. When option specifies a
// move, the call ensures that the specified files at external_file_path are
// deleted on successful return and files are not modified on any error
// return.
// On error return, column family handle returned will be nullptr.
// ColumnFamily will be present on successful return and will not be present
// on error return. ColumnFamily may be present on any crash during this call.
virtual Status CreateColumnFamilyWithImport(
const ColumnFamilyOptions& options, const std::string& column_family_name,
const ImportColumnFamilyOptions& import_options,
const ExportImportFilesMetaData& metadata,
ColumnFamilyHandle** handle);
Internally, this API creates a new CF, parses all the sst files and adds it to the specified column family, at the same level and with same sequence number as in the metadata. Also performs safety checks with respect to overlaps between the sst files being imported.
If incoming sequence number is higher than current local sequence number, local sequence
number is updated to reflect this.
Note, as the sst files is are being moved across Column Families, Column Family name in sst file
will no longer match the actual column family on destination DB. The API does not modify Column
Family name or id in the sst files being imported.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5495
Differential Revision: D16018881
fbshipit-source-id: 9ae2251025d5916d35a9fc4ea4d6707f6be16ff9
Summary:
1. Cleanup WAL trash files on open
2. Don't apply deletion rate limit if WAL dir is different from db dir
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5520
Test Plan: Add new unit tests and make check
Differential Revision: D16096750
Pulled By: anand1976
fbshipit-source-id: 6f07858ad864b754b711db416f0389c45ede599b
Summary:
This is to prevent bg flush thread from unrefing and deleting the cfd that has been dropped by a concurrent thread.
Before RocksDB calls `DBImpl::WaitForFlushMemTables`, we should increase the refcount of each `ColumnFamilyData` so that its ref count will not drop to 0 even if the column family is dropped by another thread. Otherwise the bg flush thread can deref the cfd and deletes it, causing a segfault in `WaitForFlushMemtables` upon accessing `cfd`.
Test plan (on devserver):
```
$make clean && COMPILE_WITH_ASAN=1 make -j32
$make check
```
All unit tests must pass.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5513
Differential Revision: D16062898
Pulled By: riversand963
fbshipit-source-id: 37dc511f1dc99f036d0201bbd7f0a8f5677c763d
Summary:
WAL records RocksDB writes to all column families. When user flushes a a column family, the old WAL will not accept new writes but cannot be deleted yet because it may still contain live data for other column families. (See https://github.com/facebook/rocksdb/wiki/Write-Ahead-Log#life-cycle-of-a-wal for detailed explanation)
Because of this, if there is a column family that receive very infrequent writes and no manual flush is called for it, it could prevent a lot of WALs from being deleted. PR https://github.com/facebook/rocksdb/pull/5046 introduced persistent stats column family which is a good example of such column families. Depending on the config, it may have long intervals between writes, and user is unaware of it which makes it difficult to call manual flush for it.
This PR addresses the problem for persistent stats column family by forcing a flush for persistent stats column family when 1) another column family is flushed 2) persistent stats column family's log number is the smallest among all column families, this way persistent stats column family will keep advancing its log number when necessary, allowing RocksDB to delete old WAL files.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5509
Differential Revision: D16045896
Pulled By: miasantreble
fbshipit-source-id: 286837b633e988417f0096ff38384742d3b40ef4
Summary:
It it not safe to assume application had sync the SST file before ingest it into DB. Also the directory to put the ingested file needs to be fsync, otherwise the file can be lost. For integrity of RocksDB we need to sync the ingested file and directory before apply the change to manifest.
Also syncing after writing global sequence when write_global_seqno=true was removed in https://github.com/facebook/rocksdb/issues/4172. Adding it back.
Fixes https://github.com/facebook/rocksdb/issues/5287.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5435
Test Plan:
Test ingest file with ldb command and observe fsync/fdatasync in strace output. Tried both move_files=true and move_files=false.
https://gist.github.com/yiwu-arbug/650a4023f57979056d83485fa863bef9
More test suggestions are welcome.
Differential Revision: D15941675
Pulled By: riversand963
fbshipit-source-id: 389533f3923065a96df2cdde23ff4724a1810d78
Summary:
This PR adds more callers for table readers. These information are only used for block cache analysis so that we can know which caller accesses a block.
1. It renames the BlockCacheLookupCaller to TableReaderCaller as passing the caller from upstream requires changes to table_reader.h and TableReaderCaller is a more appropriate name.
2. It adds more table reader callers in table/table_reader_caller.h, e.g., kCompactionRefill, kExternalSSTIngestion, and kBuildTable.
This PR is long as it requires modification of interfaces in table_reader.h, e.g., NewIterator.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5454
Test Plan: make clean && COMPILE_WITH_ASAN=1 make check -j32.
Differential Revision: D15819451
Pulled By: HaoyuHuang
fbshipit-source-id: b6caa704c8fb96ddd15b9a934b7e7ea87f88092d
Summary:
While the secondary is replaying after the primary, the primary may switch to a new MANIFEST. The secondary is already able to detect and follow the primary to the new MANIFEST. However, the current implementation has a bug, described as follows.
The new MANIFEST's first records have been generated by VersionSet::WriteSnapshot to describe the current state of the column families and the db as of the MANIFEST creation. Since the secondary instance has already finished recovering upon start, there is no need for the secondary to process these records. Actually, if the secondary were to replay these records, the secondary may end up adding the same SST files **again** to each column family, causing consistency checks done by VersionBuilder to fail. Therefore, we record the number of records to skip at the beginning of the new MANIFEST and ignore them.
Test plan (on dev server)
```
$make clean && make -j32 all
$./db_secondary_test
```
All existing unit tests must pass as well.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5472
Differential Revision: D15866771
Pulled By: riversand963
fbshipit-source-id: a1eec4837fb2ad13059398efb0f437e74fd53bed
Summary:
`DBImplSecondary` calls `CheckConsistency()` during open. In the past, `DBImplSecondary` did not override this function thus `DBImpl::CheckConsistency()` is called.
The following can happen. The secondary instance is performing consistency check which calls `GetFileSize(file_path)` but the file at `file_path` is deleted by the primary instance. `DBImpl::CheckConsistency` does not account for this and fails the consistency check. This is undesirable. The solution is that, we call `DBImpl::CheckConsistency()` first. If it passes, then we are good. If not, we give it a second chance and handles the case of file(s) being deleted.
Test plan (on dev server):
```
$make clean && make -j20 all
$./db_secondary_test
```
All other existing unit tests must pass as well.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5469
Differential Revision: D15861845
Pulled By: riversand963
fbshipit-source-id: 507d72392508caed3cd003bb2e2aa43f993dd597
Summary:
This PR continues the work in https://github.com/facebook/rocksdb/pull/4748 and https://github.com/facebook/rocksdb/pull/4535 by adding a new DBOption `persist_stats_to_disk` which instructs RocksDB to persist stats history to RocksDB itself. When statistics is enabled, and both options `stats_persist_period_sec` and `persist_stats_to_disk` are set, RocksDB will periodically write stats to a built-in column family in the following form: key -> (timestamp in microseconds)#(stats name), value -> stats value. The existing API `GetStatsHistory` will detect the current value of `persist_stats_to_disk` and either read from in-memory data structure or from the hidden column family on disk.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5046
Differential Revision: D15863138
Pulled By: miasantreble
fbshipit-source-id: bb82abdb3f2ca581aa42531734ac799f113e931b
Summary:
It seems like CF Options are not properly validated when creating a new column family with `CreateColumnFamily` API; only a selected few checks are done. Calling `ColumnFamilyData::ValidateOptions`, which is the single source for all CFOptions validations, will help fix this. (`ColumnFamilyData::ValidateOptions` is already called at the time of `DB::Open`).
**Test Plan:**
Added a new test: `DBTest.CreateColumnFamilyShouldFailOnIncompatibleOptions`
```
TEST_TMPDIR=/dev/shm ./db_test --gtest_filter=DBTest.CreateColumnFamilyShouldFailOnIncompatibleOptions
```
Also ran gtest-parallel to make sure the new test is not flaky.
```
TEST_TMPDIR=/dev/shm ~/gtest-parallel/gtest-parallel ./db_test --gtest_filter=DBTest.CreateColumnFamilyShouldFailOnIncompatibleOptions --repeat=10000
[10000/10000] DBTest.CreateColumnFamilyShouldFailOnIncompatibleOptions (15 ms)
```
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5453
Differential Revision: D15816851
Pulled By: sagar0
fbshipit-source-id: 9e702b9850f5c4a7e0ef8d39e1e6f9b81e7fe1e5
Summary:
This PR integrates the block cache tracer class into db_impl.cc.
db_impl.cc contains a member variable of AtomicBlockCacheTraceWriter class and passes its reference to the block_based_table_reader.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5433
Differential Revision: D15728016
Pulled By: HaoyuHuang
fbshipit-source-id: 23d5659e8c82d556833dcc1a5558aac8c1f7db71
Summary:
In secondary mode, it is possible that the secondary lists the primary's WAL
directory, finds a WAL and tries to open it. It is possible that the primary
deletes the WAL after secondary listing dir but before the secondary opening
it. Then the secondary will fail to open the WAL file with a PathNotFound
status. In this case, we can return OK without replaying WAL and optionally
replay more MANIFEST.
Test Plan (on my dev machine):
Without this PR, the following will fail several times out of 100 runs.
```
~/gtest-parallel/gtest-parallel -r 100 -w 16 ./db_secondary_test --gtest_filter=DBSecondaryTest.SwitchToNewManifestDuringOpen
```
With this PR, the above should always succeed.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5323
Differential Revision: D15763878
Pulled By: riversand963
fbshipit-source-id: c7164fa7cb8d9001abc258b6a2dc93613e4f38ff
Summary:
If a memtable definitely covers a key, there isn't a need to check older memtables.
We can skip them by checking the earliest sequence number.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/4941
Differential Revision: D13932666
fbshipit-source-id: b9d52f234b8ad9dd3bf6547645cd457175a3ca9b
Summary:
To avoid deadlock mutex_ should never be acquired before log_write_mutex_. The patch documents that and also fixes one case in ::FlushWAL that acquires mutex_ through ::WriteStatusCheck when it already holds lock on log_write_mutex_.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5437
Differential Revision: D15749722
Pulled By: maysamyabandeh
fbshipit-source-id: f57b69c44b4b80cc6d7ddf3d3fdf4a9eb5a5a45a
Summary:
Use `CreateLoggerFromOptions` function to reduce code duplication.
Test plan (on my machine)
```
$make clean && make -j32 db_secondary_test
$KEEP_DB=1 ./db_secondary_test
```
Verify all info logs of the secondary instance are properly logged.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5427
Differential Revision: D15748922
Pulled By: riversand963
fbshipit-source-id: bad7261df1b8373efc504f141efc7871e375a311
Summary:
BlockCacheLookupContext only contains the caller for now.
We will trace block accesses at five places:
1. BlockBasedTable::GetFilter.
2. BlockBasedTable::GetUncompressedDict.
3. BlockBasedTable::MaybeReadAndLoadToCache. (To trace access on data, index, and range deletion block.)
4. BlockBasedTable::Get. (To trace the referenced key and whether the referenced key exists in a fetched data block.)
5. BlockBasedTable::MultiGet. (To trace the referenced key and whether the referenced key exists in a fetched data block.)
We create the context at:
1. BlockBasedTable::Get. (kUserGet)
2. BlockBasedTable::MultiGet. (kUserMGet)
3. BlockBasedTable::NewIterator. (either kUserIterator, kCompaction, or external SST ingestion calls this function.)
4. BlockBasedTable::Open. (kPrefetch)
5. Index/Filter::CacheDependencies. (kPrefetch)
6. BlockBasedTable::ApproximateOffsetOf. (kCompaction or kUserApproximateSize).
I loaded 1 million key-value pairs into the database and ran the readrandom benchmark with a single thread. I gave the block cache 10 GB to make sure all reads hit the block cache after warmup. The throughput is comparable.
Throughput of this PR: 231334 ops/s.
Throughput of the master branch: 238428 ops/s.
Experiment setup:
RocksDB: version 6.2
Date: Mon Jun 10 10:42:51 2019
CPU: 24 * Intel Core Processor (Skylake)
CPUCache: 16384 KB
Keys: 20 bytes each
Values: 100 bytes each (100 bytes after compression)
Entries: 1000000
Prefix: 20 bytes
Keys per prefix: 0
RawSize: 114.4 MB (estimated)
FileSize: 114.4 MB (estimated)
Write rate: 0 bytes/second
Read rate: 0 ops/second
Compression: NoCompression
Compression sampling rate: 0
Memtablerep: skip_list
Perf Level: 1
Load command: ./db_bench --benchmarks="fillseq" --key_size=20 --prefix_size=20 --keys_per_prefix=0 --value_size=100 --statistics --cache_index_and_filter_blocks --cache_size=10737418240 --disable_auto_compactions=1 --disable_wal=1 --compression_type=none --min_level_to_compress=-1 --compression_ratio=1 --num=1000000
Run command: ./db_bench --benchmarks="readrandom,stats" --use_existing_db --threads=1 --duration=120 --key_size=20 --prefix_size=20 --keys_per_prefix=0 --value_size=100 --statistics --cache_index_and_filter_blocks --cache_size=10737418240 --disable_auto_compactions=1 --disable_wal=1 --compression_type=none --min_level_to_compress=-1 --compression_ratio=1 --num=1000000 --duration=120
TODOs:
1. Create a caller for external SST file ingestion and differentiate the callers for iterator.
2. Integrate tracer to trace block cache accesses.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5421
Differential Revision: D15704258
Pulled By: HaoyuHuang
fbshipit-source-id: 4aa8a55f8cb1576ffb367bfa3186a91d8f06d93a
Summary:
In regular RocksDB instance, `MemTable::earliest_seqno_` is "db sequence number at the time of creation". However, we cannot use the db sequence number to set the value of `MemTable::earliest_seqno_` for secondary instance, i.e. `DBImplSecondary` due to the logic of MANIFEST and WAL replay.
When replaying the log files of the primary, the secondary instance first replays MANIFEST and updates the db sequence number if necessary. Next, the secondary replays WAL files, creates new memtables if necessary and inserts key-value pairs into memtables. The following can occur when the db has two or more column families.
Assume the db has column family "default" and "cf1". At a certain in time, both "default" and "cf1" have data in memtables.
1. Primary triggers a flush and flushes "cf1". "default" is **not** flushed.
2. Secondary replays the MANIFEST updates its db sequence number to the latest value learned from the MANIFEST.
3. Secondary starts to replay WAL that contains the writes to "default". It is possible that the write batches' sequence numbers are smaller than the db sequence number. In this case, these write batches will be skipped, and these updates will not be visible to reader until "default" is later flushed.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5413
Differential Revision: D15637407
Pulled By: riversand963
fbshipit-source-id: 3de3fe35cfc6f1b9f844f3f926f0df29717b6580
Summary:
The patch reduces the contention over prepared_mutex_ using these techniques:
1) Move ::RemovePrepared() to be called from the commit callback when we have two write queues.
2) Use two separate mutex for PreparedHeap, one prepared_mutex_ needed for ::RemovePrepared, and one ::push_pop_mutex() needed for ::AddPrepared(). Given that we call ::AddPrepared only from the first write queue and ::RemovePrepared mostly from the 2nd, this will result into each the two write queues not competing with each other over a single mutex. ::RemovePrepared might occasionally need to acquire ::push_pop_mutex() if ::erase() ends up with calling ::pop()
3) Acquire ::push_pop_mutex() on the first callback of the write queue and release it on the last.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5420
Differential Revision: D15741985
Pulled By: maysamyabandeh
fbshipit-source-id: 84ce8016007e88bb6e10da5760ba1f0d26347735
Summary:
I'm not able to prove it, but the stress test failure may be caused by the following sequence of events -
1. Crash db_stress while writing the log file. This should result in a corrupted WAL.
2. Run db_stress with recycle_log_file_num=1. Crash during recovery immediately after writing manifest and updating the current file. The old log from the previous run is left behind, but the memtable would have been flushed during recovery and the CF log number will point to the newer log
3. Run db_stress with recycle_log_file_num=0. During recovery, the old log file will be processed and the corruption will be detected. Since the CF has moved ahead, we get the "SST file is ahead of WAL" error
Test -
1. stress_crash
2. make check
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5412
Differential Revision: D15699120
Pulled By: anand1976
fbshipit-source-id: 9092ce81e7c4a0b4b4e66560c23ea4812a4d9cbe
Summary:
When using `PRIu64` type of printf specifier, current code base does the following:
```
#ifndef __STDC_FORMAT_MACROS
#define __STDC_FORMAT_MACROS
#endif
#include <inttypes.h>
```
However, this can be simplified to
```
#include <cinttypes>
```
as long as flag `-std=c++11` is used.
This should solve issues like https://github.com/facebook/rocksdb/issues/5159
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5402
Differential Revision: D15701195
Pulled By: miasantreble
fbshipit-source-id: 6dac0a05f52aadb55e9728038599d3d2e4b59d03
Summary:
It's useful to be able to (optionally) associate key-value pairs with user-provided timestamps. This PR is an early effort towards this goal and continues the work of facebook#4942. A suite of new unit tests exist in DBBasicTestWithTimestampWithParam. Support for timestamp requires the user to provide timestamp as a slice in `ReadOptions` and `WriteOptions`. All timestamps of the same database must share the same length, format, etc. The format of the timestamp is the same throughout the same database, and the user is responsible for providing a comparator function (Comparator) to order the <key, timestamp> tuples. Once created, the format and length of the timestamp cannot change (at least for now).
Test plan (on devserver):
```
$COMPILE_WITH_ASAN=1 make -j32 all
$./db_basic_test --gtest_filter=Timestamp/DBBasicTestWithTimestampWithParam.PutAndGet/*
$make check
```
All tests must pass.
We also run the following db_bench tests to verify whether there is regression on Get/Put while timestamp is not enabled.
```
$TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=fillseq,readrandom -num=1000000
$TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=fillrandom -num=1000000
```
Repeat for 6 times for both versions.
Results are as follows:
```
| | readrandom | fillrandom |
| master | 16.77 MB/s | 47.05 MB/s |
| PR5079 | 16.44 MB/s | 47.03 MB/s |
```
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5079
Differential Revision: D15132946
Pulled By: riversand963
fbshipit-source-id: 833a0d657eac21182f0f206c910a6438154c742c
Summary:
With this commit, RocksDB secondary instance respects atomic groups in version edits.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5411
Differential Revision: D15617512
Pulled By: HaoyuHuang
fbshipit-source-id: 913f4ede391d772dcaf5649e3cd2099fa292d120
Summary:
The PR #5275 separated the column dropped and shutdown status codes. However, there were a couple of places in compaction where this change ended up treating a ShutdownInProgress() error as a real error and set bg_error. This caused MyRocks unit test to fail due to WAL writes during shutdown returning this error. Fix it by ignoring the shutdown status during compaction.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5400
Differential Revision: D15611680
Pulled By: anand1976
fbshipit-source-id: c602e97840e3ae24eb420d61e0ce95d3e6258632
Summary:
Currently we validate options in DB::Open. However the validation step is missing when options are dynamically updated in ::SetOptions. The patch fixes that.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5368
Differential Revision: D15540101
Pulled By: maysamyabandeh
fbshipit-source-id: d27bbffd8f0252d1b50bcf59e0a70a278ed937f4
Summary:
util/ means for lower level libraries. trace_replay is highly integrated to DB and sometimes call DB. Move it out to a separate directory.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5376
Differential Revision: D15550938
Pulled By: siying
fbshipit-source-id: f46dce5ceffdc05a73f26379c7bb1b79ebe6c207
Summary:
Many logging related source files are under util/. It will be more structured if they are together.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5387
Differential Revision: D15579036
Pulled By: siying
fbshipit-source-id: 3850134ed50b8c0bb40a0c8ae1f184fa4081303f
Manuel Ung