Skip to content

Latest commit

 

History

History
314 lines (239 loc) · 16.5 KB

tiproxy-performance-test.md

File metadata and controls

314 lines (239 loc) · 16.5 KB
Raw
title summary
TiProxy Performance Test Report
Learn the performance of TiProxy and the comparison with HAProxy.

TiProxy Performance Test Report

This report tests the performance of TiProxy in the OLTP scenario of Sysbench and compares it with HAProxy.

The results are as follows:

  • The QPS upper limit of TiProxy is affected by the type of workload. Under the basic workloads of Sysbench and the same CPU usage, the QPS of TiProxy is about 25% lower than that of HAProxy.
  • The number of TiDB server instances that TiProxy can hold varies according to the type of workload. Under the basic workloads of Sysbench, a TiProxy can hold 5 to 12 TiDB server instances of the same model.
  • The row number of the query result set has a significant impact on the QPS of TiProxy, and the impact is the same as that of HAProxy.
  • The performance of TiProxy increases almost linearly with the number of vCPUs. Therefore, increasing the number of vCPUs can effectively improve the QPS upper limit.
  • The number of long connections and the frequency of creating short connections have minimal impact on the QPS of TiProxy.

Test environment

Hardware configuration

Service Machine type CPU model Instance count
TiProxy 4C8G Intel(R) Xeon(R) Silver 4214R CPU @ 2.40GHz 1
HAProxy 4C8G Intel(R) Xeon(R) Silver 4214R CPU @ 2.40GHz 1
PD 4C8G Intel(R) Xeon(R) Silver 4214R CPU @ 2.40GHz 3
TiDB 8C16G Intel(R) Xeon(R) Silver 4214R CPU @ 2.40GHz 8
TiKV 8C16G Intel(R) Xeon(R) Silver 4214R CPU @ 2.40GHz 8
Sysbench 8C16G Intel(R) Xeon(R) Silver 4214R CPU @ 2.40GHz 1

Software

Service Software version
TiProxy v1.0.0
HAProxy 2.9.0
PD v8.0.0
TiDB v8.0.0
TiKV v8.0.0
Sysbench 1.0.17

Configuration

TiProxy configuration

In the test, TLS is not enabled between the client and TiProxy, or between TiProxy and TiDB server.

proxy.conn-buffer-size: 131072

HAProxy configuration - haproxy.cfg file

global                                      # Global configuration.
    log         127.0.0.1 local2            # Specifies the global syslog server. You can define up to two.
    pidfile     /var/run/haproxy.pid        # Write the PID of the HAProxy process to pidfile.
    maxconn     4096                        # The maximum number of concurrent connections that a single HAProxy process can accept, equivalent to the command line parameter "-n".
    nbthread    4                           # The maximum number of threads. The upper limit of the number of threads is the same as the number of CPUs.
    user        haproxy                     # Same as UID parameter.
    group       haproxy                     # Same as GID parameter. A dedicated user group is recommended.
    daemon                                  # Run HAProxy as a daemon in the background, which is equivalent to the command line parameter "-D". You can also disable it with the "-db" parameter on the command line.
    stats socket /var/lib/haproxy/stats     # The location where the statistics are saved.

defaults                                    # Default configuration.
    log global                              # Inherit the settings of the global configuration section.
    retries 2                               # The maximum number of times to try to connect to the upstream server. If the number of retries exceeds this value, the backend server is considered unavailable.
    timeout connect  2s                     # The timeout period for HAProxy to connect to the backend server. If the server is located on the same LAN as HAProxy, set it to a shorter time.
    timeout client 30000s                   # The timeout period for the client to be inactive after the data transmission is completed.
    timeout server 30000s                   # The timeout period for the server to be inactive.

listen tidb-cluster                         # Configure database load balancing.
    bind 0.0.0.0:3390                       # Floating IP address and listening port.
    mode tcp                                # HAProxy uses layer 4, the transport layer.
    balance leastconn                      # The server with the least number of connections receives the connection first. `leastconn` is recommended for long session services, such as LDAP, SQL, and TSE, rather than short session protocols, such as HTTP. This algorithm is dynamic, and the server weight is adjusted during operation for slow-start servers.
    server tidb-1 10.9.18.229:4000 check inter 2000 rise 2 fall 3      # Detects port 4000 at a frequency of once every 2000 milliseconds. If it is detected as successful twice, the server is considered available; if it is detected as failed three times, the server is considered unavailable.
    server tidb-2 10.9.39.208:4000 check inter 2000 rise 2 fall 3
    server tidb-3 10.9.64.166:4000 check inter 2000 rise 2 fall 3

Basic workload test

Test plan

This test aims to compare the QPS of TiProxy and HAProxy under four types of workloads: point select, read only, write only, and read write. Each type of workload is tested with different concurrency to compare the QPS of TiProxy and HAProxy.

The following command is used to perform the test:

sysbench $testname \
    --threads=$threads \
    --time=1200 \
    --report-interval=10 \
    --rand-type=uniform \
    --db-driver=mysql \
    --mysql-db=sbtest \
    --mysql-host=$host \
    --mysql-port=$port \
    run --tables=32 --table-size=1000000

Point Select

TiProxy test results:

Threads QPS Avg latency (ms) P95 latency (ms) TiProxy CPU usage TiDB overall CPU usage
20 41273 0.48 0.64 190% 900%
50 100255 0.50 0.62 330% 1900%
100 137688 0.73 1.01 400% 2600%

HAProxy test results:

Threads QPS Avg latency (ms) P95 latency (ms) HAProxy CPU usage TiDB overall CPU usage
20 44833 0.45 0.61 140% 1000%
50 103631 0.48 0.61 270% 2100%
100 163069 0.61 0.77 360% 3100%

Read Only

TiProxy test results:

Threads QPS Avg latency (ms) P95 latency (ms) TiProxy CPU usage TiDB overall CPU usage
50 72076 11.09 12.75 290% 2500%
100 109704 14.58 17.63 370% 3800%
200 117519 27.21 32.53 400% 4100%

HAProxy test results:

Threads QPS Avg latency (ms) P95 latency (ms) HAProxy CPU usage TiDB overall CPU usage
50 75760 10.56 12.08 250% 2600%
100 121730 13.14 15.83 350% 4200%
200 131712 24.27 30.26 370% 4500%

Write Only

TiProxy test results:

Threads QPS Avg latency (ms) P95 latency (ms) TiProxy CPU usage TiDB overall CPU usage
100 81957 7.32 10.27 290% 3900%
300 103040 17.45 31.37 330% 4700%
500 104869 28.59 52.89 340% 4800%

HAProxy test results:

Threads QPS Avg latency (ms) P95 latency (ms) HAProxy CPU usage TiDB overall CPU usage
100 81708 7.34 10.65 240% 3700%
300 106008 16.95 31.37 320% 4800%
500 122369 24.45 47.47 350% 5300%

Read Write

TiProxy test results:

Threads QPS Avg latency (ms) P95 latency (ms) TiProxy CPU usage TiDB overall CPU usage
50 58571 17.07 19.65 250% 2600%
100 88432 22.60 29.19 330% 3900%
200 108758 36.73 51.94 380% 4800%

HAProxy test results:

Threads QPS Avg latency (ms) P95 latency (ms) HAProxy CPU usage TiDB overall CPU usage
50 61226 16.33 19.65 190% 2800%
100 96569 20.70 26.68 290% 4100%
200 120163 31.28 49.21 340% 5200%

Result set test

Test plan

This test aims to compare the performance of TiProxy and HAProxy under different result set row numbers. This test uses 100 concurrency, and compares the QPS of TiProxy and HAProxy with result set row numbers of 10, 100, 1000, and 10000.

The following command is used to perform the test:

sysbench oltp_read_only \
    --threads=100 \
    --time=1200 \
    --report-interval=10 \
    --rand-type=uniform \
    --db-driver=mysql \
    --mysql-db=sbtest \
    --mysql-host=$host \
    --mysql-port=$port \
    --skip_trx=true \
    --point_selects=0 \
    --sum_ranges=0 \
    --order_ranges=0 \
    --distinct_ranges=0 \
    --simple_ranges=1 \
    --range_size=$range_size
    run --tables=32 --table-size=1000000

Test results

TiProxy test results:

Range size QPS Avg latency (ms) P95 latency (ms) TiProxy CPU usage TiDB overall CPU usage Inbound network (MiB/s) Outbound network (MiB/s)
10 80157 1.25 1.61 340% 2600% 140 140
100 55936 1.79 2.43 370% 2800% 820 820
1000 10313 9.69 13.70 310% 1500% 1370 1370
10000 1064 93.88 142.39 250% 600% 1430 1430

HAProxy test results:

Range size QPS Avg latency (ms) P95 latency (ms) HAProxy CPU usage TiDB overall CPU usage Inbound network (MiB/s) Outbound network (MiB/s)
10 94376 1.06 1.30 250% 4000% 150 150
100 70129 1.42 1.76 270% 3300% 890 890
1000 9501 11.18 14.73 240% 1500% 1180 1180
10000 955 104.61 320.17 180% 1200% 1200 1200

Scalability test

Test plan

This test aims to verify that the performance of TiProxy is proportional to its specifications, to ensure that upgrading the specifications of TiProxy can improve its QPS upper limit. This test uses TiProxy instances with different vCPU numbers and concurrency to compare the QPS.

The following command is used to perform the test:

sysbench oltp_point_select \
    --threads=$threads \
    --time=1200 \
    --report-interval=10 \
    --rand-type=uniform \
    --db-driver=mysql \
    --mysql-db=sbtest \
    --mysql-host=$host \
    --mysql-port=$port \
    run --tables=32 --table-size=1000000

Test results

vCPU Threads QPS Avg latency (ms) P95 latency (ms) TiProxy CPU usage TiDB overall CPU usage
2 40 58508 0.68 0.97 190% 1200%
4 80 104890 0.76 1.16 390% 2000%
6 120 155520 0.77 1.14 590% 2900%
8 160 202134 0.79 1.18 800% 3900%

Long connection test

Test plan

This test aims to verify that a large number of idle connections have minimal impact on the QPS when the client uses long connections. This test creates 5000, 10000, and 15000 idle long connections, and then executes sysbench.

This test uses the default value for the conn-buffer-size configuration:

proxy.conn-buffer-size: 32768

Use the following command to perform the test:

sysbench oltp_point_select \
    --threads=50 \
    --time=1200 \
    --report-interval=10 \
    --rand-type=uniform \
    --db-driver=mysql \
    --mysql-db=sbtest \
    --mysql-host=$host \
    --mysql-port=$port \
    run --tables=32 --table-size=1000000

Test results

Connection count QPS Avg latency (ms) P95 latency (ms) TiProxy CPU usage TiProxy memory usage (MB) TiDB overall CPU usage
5000 96620 0.52 0.64 330% 920 1800%
10000 96143 0.52 0.65 330% 1710 1800%
15000 96048 0.52 0.65 330% 2570 1900%

Short connection test

Test plan

This test aims to verify that frequent creation and destruction of connections have minimal impact on the QPS when the client uses short connections. This test starts another client program to create and disconnect 100, 200, and 300 short connections per second while executing sysbench.

Use the following command to perform the test:

sysbench oltp_point_select \
    --threads=50 \
    --time=1200 \
    --report-interval=10 \
    --rand-type=uniform \
    --db-driver=mysql \
    --mysql-db=sbtest \
    --mysql-host=$host \
    --mysql-port=$port \
    run --tables=32 --table-size=1000000

Test results

New connections per second QPS Avg latency (ms) P95 latency (ms) TiProxy CPU usage TiDB overall CPU usage
100 95597 0.52 0.65 330% 1800%
200 94692 0.53 0.67 330% 1800%
300 94102 0.53 0.68 330% 1900%