0

I have the following setup.

2 x gce n1.standard instances running nginx/php5-fpm

1 x cloud SQL d8 instance with both GCE instances connected to it.

The front end servers are running a stripped down version of osCommerce-2.3.3.4 with NO admin section on the front end. Its just the catalog portion of osC.

I ran a load test with load impact, and the site is unusable at around 50-100 users. I am only looking to support at absolute max 500/users at any give time. We average around 130-170 on our current server.

I am not looking for a complete explanation but any helpful places to check, things to try, stuff to read, I just need a direction to go in to get this cloud platform working like we wanted.

Thanks in advance.

Justin W. Johns
  • 119
  • 1
  • 8
  • Could you tell more about the environment e.g., 1) What are the types of queries; are they mostly reads / writes ? 2) What is the QPS that you expect ? 3) Are you using any sort caching ? 4) Is the data different in some way (size / indexe(s) / any other particular characteristic of the data) ? – Aditya Mone Feb 11 '14 at 19:04
  • mainly reads, only writes would be session values. QPS/s, I did not have a target, I just wanted to help our site under load. I am not familiar with sort caching so no. The data is from a factory installation of oscommerce-2.3.3.4, there really it not much data there at all. I think it is the traffic bogging it down, although I am unsure. – Justin W. Johns Feb 11 '14 at 19:19
  • I solved my problem with better configuration of nginx itself. I only had 1 worker process for starters. http://lowlatencyweb.wordpress.com/2012/03/20/500000-requestssec-modern-http-servers-are-fast/ is what I found looking for solutions. – Justin W. Johns Feb 11 '14 at 22:51

0 Answers0