XF 2.3 Install / Upgrade with Auto-Scaling

[Spectral-Brain]

I am using AWS and want to setup auto-scaling for when our forums are needing to scale up due to more usage. I've seen in this forum that others have auto scaling so want to understand how they have set this up, and how they upgrade.

The solution I am thinking for auto scaling is to setup an ec2 instance with xenforo, then create an AMI template from that ec2 instance which will be used for auto scaling. I would move the data and internal_data to s3 and use an RDS database. But when it is time to upgrade, I'm not sure how to go about that.

How do you guys that utilize auto scaling do it?

 

[frm]

Following as I'd like to auto-scale as well, either that or load balancing by popping up a duplicate instance in a heavily trafficked region until it's no longer needed.

I just went through manual scaling myself yesterday, but there's still a problem with configuration after that's done. I expanded Elasticsearch memory from 1GB (yes, I managed to get it to run on 2GB) to 4GB. So, there still may be manual settings that you need to go over (such as editing caching file sizes, etc., unless auto-scaling can tell Apache/nginx/Redis/Elasticsearch/PHP to update how much to cache/bump the max memory limits on scale).

 

[philmckrackon]

I don't use AWS but thought I would add some info. I don't have nearly the traffic needed for auto scaling.
Good job on ES, finding the sweet spot is key. I have ES running on 1.6G without ever having the process stop and restart during peak usage or during backups when memory lowered. Only 400k docs indexed though.


As for REDIS,my sweet spot is 256M (I cache guests). Your settings will vary.


I also just optimized Opcache using this great PHP script from Github. All the information is viewable under PHP info, this just gives a graphical view.
Thread 'Great Opcache tuning script' https://xenforo.com/community/threads/great-opcache-tuning-script.228992/