Running sysbench against volumes I'm finding the gp3 volumes are much slower. Even when I provision a volume with 16000IOPS after waiting for optimization I'm getting a cap at 1000 IOPS when monitoring through Percona PMM, New Relic CloudWatch.
Any ideas? Thanks, Jeremy
Instance:
Volumes:
Sysbench results below:
sysbench --file-total-size=15G --file-num=16 fileio prepare
sysbench --file-total-size=15G --file-num=16 --file-test-mode=rndrw --time=600 fileio run
sysbench --file-total-size=15G --file-num=16 fileio cleanup
gp3 | 9000 | 3000 | 250/s
File operations:
reads/s: 576.37
writes/s: 384.24
fsyncs/s: 153.70
Throughput:
read, MiB/s: 9.01
written, MiB/s: 6.00
General statistics:
total time: 600.0333s
total number of events: 668612
Latency (ms):
min: 0.00
avg: 0.90
max: 337.40
95th percentile: 3.89
sum: 599693.33
Threads fairness:
events (avg/stddev): 668612.0000/0.00
execution time (avg/stddev): 599.6933/0.00
gp2 | 9000 | 3000 | gp2
File operations:
reads/s: 1523.68
writes/s: 1015.79
fsyncs/s: 406.33
Throughput:
read, MiB/s: 23.81
written, MiB/s: 15.87
General statistics:
total time: 600.0064s
total number of events: 1767487
Latency (ms):
min: 0.00
avg: 0.34
max: 70.10
95th percentile: 1.06
sum: 599390.12
Threads fairness:
events (avg/stddev): 1767487.0000/0.00
execution time (avg/stddev): 599.3901/0.00
Anyone have ideas on this? I've posted on the AWS forums as well with no response.
Here are some more test results with the same volumes as above except I increased the gp3 volumes to 16000 IOPS and 1000/s throughput and still get worse results.
Pics are 2 gp3 volumes on the left and the old gp2 on the right. If there are two humps the first one is writing the sysbench test files, second is the test.
Thanks, Jeremy
My initial enthusiasm for gp3 was dampened by inferior performance to gp2, however, when I set the size, IOPS, and bandwidth the same on C5, M5, and M5A instance type I got similar performance measurements from fio
: https://github.com/axboe/fio
I'm testing on CentOS 7.8, fio-3.7-2.el7.x86_64
. Region us-east-1
.
Are you sure sysbench is the right benchmarking tool? The fileio
options don't seem greatly documented, and I can get wildly different results when changing the sync mode, flags, etc.
Your example command equivalent to this because of the default values:
sysbench --file-total-size=15G --file-num=16 --file-test-mode=rndrw --file-io-mode=sync --file-rw-ratio=1.5 --file-fsync-freq=100 --max-requests=0 --time=600 fileio run
When I run a different sysbench command, notably using direct
access and fsync=off
:
sysbench fileio --file-total-size=15G --file-test-mode=rndrw --time=600 --max-requests=0 --file-io-mode=async --file-extra-flags=direct --file-num=16 --file-rw-ratio=1 --file-fsync-freq=0 run
I get equivalent performance from gp2 and gp3. (Actually, gp3 is slightly better).
Here's a summary of my fio
data, using the command
fio --rw=TEST --direct=1 --ioengine=libaio --bs=16k --numjobs=8 --size=1G --group_reporting
with TEST= {read,write,randread,randwrite,randrw}
gp2:
gp2-randread-16k.out: read: IOPS=3081, BW=48.2MiB/s (50.5MB/s)(8192MiB/170117msec)
gp2-randrw-16k.out: read: IOPS=1540, BW=24.1MiB/s (25.2MB/s)(4095MiB/170121msec)
gp2-randrw-16k.out: write: IOPS=1541, BW=24.1MiB/s (25.2MB/s)(4097MiB/170121msec)
gp2-randwrite-16k.out: write: IOPS=3081, BW=48.2MiB/s (50.5MB/s)(8192MiB/170116msec)
gp2-read-16k.out: read: IOPS=3090, BW=48.3MiB/s (50.6MB/s)(8192MiB/169642msec)
gp2-write-16k.out: write: IOPS=3081, BW=48.2MiB/s (50.5MB/s)(8192MiB/170115msec)
gp3:
gp3-randread-16k.out: read: IOPS=3080, BW=48.1MiB/s (50.5MB/s)(8192MiB/170188msec)
gp3-randrw-16k.out: read: IOPS=1539, BW=24.1MiB/s (25.2MB/s)(4095MiB/170238msec)
gp3-randrw-16k.out: write: IOPS=1540, BW=24.1MiB/s (25.2MB/s)(4097MiB/170238msec)
gp3-randwrite-16k.out: write: IOPS=3082, BW=48.2MiB/s (50.5MB/s)(8192MiB/170073msec)
gp3-read-16k.out: read: IOPS=3089, BW=48.3MiB/s (50.6MB/s)(8192MiB/169719msec)
gp3-write-16k.out: write: IOPS=3083, BW=48.2MiB/s (50.5MB/s)(8192MiB/170053msec)
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.