Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


6x 512mb(LB-LEMP stack) & 1x 1gb MySQL -OR- 1x 4gb stack.
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

6x 512mb(LB-LEMP stack) & 1x 1gb MySQL -OR- 1x 4gb stack.

Ok, so I'm looking for some opinions (and why exactly, outside of LBing) on a preferred setup given the specs in the title.

Stack I'm thinking would be: HAProxy, Varnish, 2x nginx, 2x php-fpm, 1x MySQL

I'm looking to have a solid foundation for a new site but I'm curious if it's going a little overboard with getting everything load-balanced, varnish'd, etc all on separate servers, especially for a fresh site or... put it all one one server.

Also, when using gluster would it be a terrible idea to put the master on the box with varnish (in the case of my above example stack) or would it be a better idea to put the master on a box by itself?

Comments

  • What is it for? One website? How much traffic do you expect?

    I would presonally just install apache and not bother with all the extra complexity. But then again i don't have so much time to waste :)

    Thanked by 10xdragon
  • You can't need that kind of complexity if you're using 512MB instances.. I'd slap it on one box and be done. Most of my sites use nginx fast cache and that is more than good enough to handle some serious traffic.

    You're creating a lot more work, and more latency by splitting it up. You can cache the crap out of anything and be fine on a smaller box.

  • What's your requirement? How much traffic? How heavy the site will be?

  • @msg7086 said:
    What's your requirement? How much traffic? How heavy the site will be?

    @mikeyur said:
    You can't need that kind of complexity if you're using 512MB instances.. I'd slap it on one box and be done. Most of my sites use nginx fast cache and that is more than good enough to handle some serious traffic.

    You're creating a lot more work, and more latency by splitting it up. You can cache the crap out of anything and be fine on a smaller box.

    I'm going to assume he's looking for redundancy.

  • I think more detail is worht it small ram machines are a little werid since you can only run a few things and can only get limited things out of them. What provider are you thinking about using for this. And which services do you plan on running.

  • @mikeyur said:
    You can't need that kind of complexity if you're using 512MB instances.. I'd slap it on one box and be done. Most of my sites use nginx fast cache and that is more than good enough to handle some serious traffic.

    You're creating a lot more work, and more latency by splitting it up. You can cache the crap out of anything and be fine on a smaller box.

    It's not a matter of complexity (I already have all my install bash scripts setup and working flawlessly.

    It's a matter of being scaleable from the start. Not putting everything on one machine and then a few months in realizing I need to load-balance and cluster servers.

    Can 512mb/1gb VPSes handle a lot of traffic, primarily a Wordpress site? Absolutely, even more when you put Varnish in front of it.

  • @msg7086 said:
    What's your requirement? How much traffic? How heavy the site will be?

    I don't really have a "requirement", I'm mainly looking for the optimal setup without blowing $100/m+ on a startup website. As for traffic? I don't think anyone can really ever answer that. Obviously if it got HUGE you wouldn't be able to handle tons of traffic, but again efficiency, scalability and optimized from the start is what I'm after.

  • pugtorpugtor Member
    edited October 2014

    @wojons said:
    I think more detail is worht it small ram machines are a little werid since you can only run a few things and can only get limited things out of them. What provider are you thinking about using for this. And which services do you plan on running.

    I only ever use DO. I've tried all the big competitors in the space and DO seems to strike a nice balance on their specs. As for services, I'm thinking HAProxy, Varnish, nginx, php-fpm (possibly HHVM since WP 3.9+ CORE is working on it), Percona and Gluster.

    As for how I would assume to be optimal setup: 512mb HAProxy, 1gb Varnish, 2x 1gb nginx/php-fpm, 2gb Percona, 512mb Gluster.

  • What type of files are you trying to sync with gluster. If its just site files and lets say not large files use are uplaoding I recomned a few things or ways to think about this.

    I would make all the all nodes juss nginx/php-fpm. if you really want varnish you can put it there also. then i would setup my database nodes . If your site files are all in a package deploy code with git or a tar to all the files . and dploy it to all the nodes and then let ur cache store the files that users are accessing if you need to have larger user fles have 2 servers 1 mysql and one diedicated to clsuter i guess and use that for file storage but i like to keep all my nodes exactly the same not including anything thatslike where a small amoutn of data is sotreid but i control that all in code.

  • Try scale up before scale out.

    If the provider does not support vip, AZ(available zone), private network; all your hard work is in vain.

Sign In or Register to comment.