db_bench reports "open error: IO error: lock hold by current process"

[Hongwei Qin]

Hi RocksDB experts,

I'm trying to run the mixgraph benchmark proposed in the FAST20 paper(https://www.usenix.org/system/files/fast20-cao_zhichao.pdf).

According to the appendix of the paper, first, I create a database with 50K kv pairs.

After that, I run the benchmark using the existing database.

Then, the db_bench tells me that "open error: IO error: lock hold by current process", which is very strange because I'm not running other benchmarks.

Same error happens after reboot.

Does anyone knows where I did wrong?

Thanks!

Regards,

Hongwei

Workflow:

[root@localhost rocksdb]# ./db_bench --benchmarks=fillrandom –perf_level=3 -use_direct_io_for_flush_and_compaction=true -use_direct_reads=true -cache_size=268435456 -key_size=48 -value_size=43 -num=50000 -db=/mnt/qhw/fast
Initializing RocksDB Options from the specified file
Initializing RocksDB Options from command-line flags
RocksDB:    version 6.15
Date:       Fri Dec  4 22:46:24 2020
CPU:        32 * Intel(R) Xeon(R) CPU E5-2650 v4 @ 2.20GHz
CPUCache:   16384 KB
Keys:       48 bytes each (+ 0 bytes user-defined timestamp)
Values:     43 bytes each (22 bytes after compression)
Entries:    50000
Prefix:    0 bytes
Keys per prefix:    0
RawSize:    4.3 MB (estimated)
FileSize:   3.3 MB (estimated)
Write rate: 0 bytes/second
Read rate: 0 ops/second
Compression: Snappy
Compression sampling rate: 0
Memtablerep: skip_list
Perf Level: 1
------------------------------------------------
Initializing RocksDB Options from the specified file
Initializing RocksDB Options from command-line flags
DB path: [/mnt/qhw/fast]
fillrandom   :       3.110 micros/op 321404 ops/sec;   27.9 MB/s

[root@localhost rocksdb]# ls /mnt/qhw
fast

[root@localhost rocksdb]# shutdown -h now

# After reboot

[root@localhost ~]# cd rocksdb/
[root@localhost rocksdb]# ./db_bench –benchmarks="mixgraph" -use_direct_io_for_flush_and_compaction=true -use_direct_reads=true -cache_size=268435456 -keyrange_num=1 -value_k=0.2615 -value_sigma=25.45 -iter_k=2.517 -iter_sigma=14.236 -mix_get_ratio=0.83 -mix_put_ratio=0.14 -mix_seek_ratio=0.03 -sine_mix_rate_interval_milliseconds=5000 -sine_a=1000 -sine_b=0.000073 -sine_d=4500 –perf_level=2 -reads=420000 -num=50000 -key_size=48 -db=/mnt/qhw/fast -use_existing_db=true
Initializing RocksDB Options from the specified file
Initializing RocksDB Options from command-line flags
RocksDB:    version 6.15
Date:       Fri Dec  4 22:49:46 2020
CPU:        32 * Intel(R) Xeon(R) CPU E5-2650 v4 @ 2.20GHz
CPUCache:   16384 KB
Keys:       48 bytes each (+ 0 bytes user-defined timestamp)
Values:     100 bytes each (50 bytes after compression)
Entries:    50000
Prefix:    0 bytes
Keys per prefix:    0
RawSize:    7.1 MB (estimated)
FileSize:   4.7 MB (estimated)
Write rate: 0 bytes/second
Read rate: 0 ops/second
Compression: Snappy
Compression sampling rate: 0
Memtablerep: skip_list
Perf Level: 1
------------------------------------------------
fillseq      : skipped (--use_existing_db is true)
Initializing RocksDB Options from the specified file
Initializing RocksDB Options from command-line flags
open error: IO error: lock hold by current process, acquire time 1607093386 acquiring thread 140305662286528: /mnt/qhw/fast/LOCK: No locks available

[曹志超]

Hi Hongwei,

Thanks for reporting the failure. We have noticed this issue several weeks ago and are working on it. We will keep you posted when it is fixed.

Thanks,

Zhichao

Hongwei Qin <glqinh...@gmail.com> 于2020年12月4日周五 上午7:27写道：

--
You received this message because you are subscribed to the Google Groups "rocksdb" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rocksdb+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/rocksdb/2a383b6d-256d-4e0e-be8f-e25c738b9880n%40googlegroups.com.

[Hongwei Qin]

Hi Zhichao, 

Thanks for your information. 

Is there another way to run the mixgraph benchmark? 

In my case,  I'm trying to use the db_bench and rocksdb to evaluate my parallelism optimization sin the ssd driver. Benchmarking would be good and real workloads would be awesome. Any suggestions? 

Thanks! 

Regards, 

Hongwei

To unsubscribe from this group and stop receiving emails from it, send an email to rocksdb+unsubscribe@googlegroups.com.

[Hongwei Qin]

Sorry for the typo.  :(

I'm trying to use the db_bench and rocksdb to evaluate my parallelism optimizations in the ssd driver.