Hi
I have setup a magento 2 host with multiple websites/domains. For testing I wrote a script that copies the production site to the testsite and does the following:
Here comes the weird part:
I use curl to test the speed of the page generation from the localhost.
curl -so /dev/null -w "%{time_connect} - %{time_starttransfer} - %{time_total} " http://www.domain.com
0.004 - 0.040 - 0.042
So the production site is loaded in 0.04x seconds. Values between 0.040 and 0.075 seconds are quite common on the production site.
Same test on the testdomain:
curl -so /dev/null -w "%{time_connect} - %{time_starttransfer} - %{time_total} " http://test.domain.com
0.004 - 1.967 - 1.980
The value of almost 2 seconds is very common on the test site.
It is the same server, running the same software.
I installed a magento 2.1.9 demoshop and did the same: So copy/rsync the demoshop1.domain.com to demoshop2.demoshop.com and these two site runs without any performance difference.
PS: I already enlarged the parameters of the opcache (which was to small to handle both sites)
Server: VM on VMWare (with SSD discs). Ubuntu 16.04 LTS, Apache2, MariaDB 10.2, PHP 7.0 with opcache
I don't know where to look any more. Does anybody have any clues?
Thom
Solved! Go to Solution.
Hmmm found it myself after 3 weeks of testing.... for some very weird reason the magento cache was enabled on the production server, but disabled on the testsite even though it as a complete copy of the production site.
Enabling the magento cache made the performance issue go away.
Thank
Thom
Hmmm found it myself after 3 weeks of testing.... for some very weird reason the magento cache was enabled on the production server, but disabled on the testsite even though it as a complete copy of the production site.
Enabling the magento cache made the performance issue go away.
Thank
Thom