Good night,
in tests the "Import large amount of data to database" getting very low score:
How to get better results? The optimization is set to 4-8GB ram in aaPanel.
MariaDB 10.6 with Redis. OLS with PHP 7.4. VPS GoogleCloud N2D 4x AMD EPYC cpu at ssd.
Stealth
Is your server high performance io?
aaPanel_Kern
Yes, from 4 selection (extreme persistent disk, normal disk...) I choose always SSD. The Contabo cheap VPS giving 6.6 total score.
aaPanel_Kern
I try it on Contabo VPS, the site and DB at 1 VPS (6 dollar/month):
Or on GoogleCloud 2 pcs of VPS, 1 is webserver, 1 is database server, same serverhouse West EU-3c (2x165 dollar/month):
"Importing large amount of data to database" is very bad only minimum 1 points, tried anything set to optimize... How set it better?
u sure its not the script idling in the background ? as from what I seen my older VPS before was 8G ram and did well for lightweight system panel... check the logs and as well the process of the system to see what's eating up the specs
aaP_hzm170
Load is 2-3%, like nothing (1 site for test). These stress-test are good, maybe try next time with Plesk (but yet not have free vps for install), I send here screenshot then.
Stealth sure I will wait when u send it over
Stealth
Did you use iotop and htop for system load before testing were there any other processes using system resources?
aaPanel_Kern
Hi, there is no load. 1 site for testing and then delete the cloud VPS. Today testing the Hestia and Plesk for this test too and add the screenshots.
Stealth
Hi, I also tested it with the tool
Server configuration: cpu 1, memory 2G, SSD, MySQL 5.7.37
Importing large amount of data to database 4.12
After modifying the configuration below 5.8 , you can also try the pre- and post-modification tests.
innodb_flush_log_at_trx_commit=2
innodb_file_per_table=1
innodb_thread_concurrency=0
innodb_buffer_pool_size=4G
innodb_buffer_pool_instances=8
innodb_log_file_size=256M
query_cache_type=1
query_cache_limit=128K
query_cache_size=64M
query_cache_min_res_unit=512
performance-schema=0
sql-mode="NO_ENGINE_SUBSTITUTION"
The innodb_flush_log_at_trx_commit=2 is very unsafe, the =0 is too, =1 is best value. Help it with speed, but big chance to loss data when hardware or software fail...
Other settings try today and send screenshot. First is Hestia (Nginx with Apache) and MariaDB (10.5?) with Redis:
(server is FREE tier Oracle Cloud ARM mobile cpu)
Stealth
Maybe this is a limitation of the cloud service provider. It is impossible to have only one server use all the performance. I feel like I don't have to waste time testing anymore.
Plesk on GoogleCloud Debian 11 VPS, ssd, Frankfurt-west 3c. Bad performance of Redis 7, because Redis can run only in Docker. That is decrease the final score. But the MariaDB 10.5 database better 3.21-3.47 yellow.
Left side PHP 7.4 and right side 8.0:
well that is something ... i have to check on my end on that ...
Okay, last picture, I tried anything set from internet, Filesystem benchmark is close to max 10 points, but the Importing larga data always bad... (DB on second VPS in 1 serverhouse)
Okay, last picture. Cheap 5β¬ VPS from Cont: details. Normal SSD (no nvme), all on 1 server, Nginx with Apache and MariaDB 10.5 with Redis 6 (not in docker but as service installed).
Tada:
6.29 point finally. Not the best, but finally more better. I tried anything on test 3pc test VPS and this is maximum. Maybe in future try the NVME SSD in Raid and 10Gbit connection and will see if better results... 
Thank you for help!
Out of curiousity which test is this? I'd like to try this too.
slimx
WordPress Hosting Benchmark tool
