This is a great starting point. What size should I use for the LB and do I need sperate servers for sphinx? John
On Thursday, October 11, 2012 12:02:41 AM UTC-4, Donovan wrote: > > Large for the MySQL is probably appropriate; I would do an nginx lb out > front on a c1.medium. If you aren't doing ssl termination a m1.small may do > fine. I think you've outgrown a t1.micro lb. > > I've tried ELBs but I didn't like losing all control of the lb, and > sometimes you need nginx's capability to solve issues, plug holes, throw up > maintenance notices, and do other optimizations. Nginx is a veritable Swiss > army knife to have at the front of your infrastructure. > > You can keep apache for your app server, I'd start with two c1.mediums for > your app servers. If your platform is stable and a known quantity on apache > switching is only likely to cause you pain. > > My guess is your sphinx servers would benefit more from ram than CPU so > I'd probably run two m1.mediums there. > > Then monitor and adjust as necessary. > > c2.mediums are going to be your processing workhorses. > > In my benchmarking for app servers a c1.medium properly tuned is better > than 3 m1.smalls. (and cheaper) > > If your serving php scalr's canned nginx, app, and MySQL roles will be > fine for you. I didn't love rds; but If I inherited a project with it I > wouldn't hate it enough to move it to non-rds instances immediately. Try > running with it until it becomes a pain point; you'll have bigger fish to > fry in the meantime. > > On Oct 10, 2012, at 2:55 PM, otherjohn <[email protected] <javascript:>> > wrote: > > I have limited information since I just took over but here is some items: > > - Site needs to be available (fail safe), we had multiple failures > before so the administrator stuck it on a m1.large with a m1.large AZ RDS > to keep it safe. > - Site has 4 million posts. The Posts table is about 2 gigs in size! > db needs to be strong enough to handle searches. > - Moving it off of vBulletin to invisionpower which runs with less > memory and can use sphinx > > Sorry this probably doesn't help much. I could probably answer particulars > if questioned. > John > > On Wednesday, October 10, 2012 5:20:36 PM UTC-4, Sebastian Stadil wrote: >> >> You'll probably want several smaller ones for the higher availability >> you'll get. Do you know what your gating items are? >> >> On Wed, Oct 10, 2012 at 2:06 PM, John H. <[email protected]> wrote: >> >>> Hi all, >>> I am looking for a setup recommendation using the AWS platform. >>> Currently we have a vBulletin site that get 1 million visits a month, >>> about 1500-2000 active visitors at peak times durring the day (12 hour >>> block), and traffic is increasing every month. >>> >>> Our current setup is a large m.1 server for this site and a large m.1 >>> RDS (multi-AZ) for the DB. >>> >>> I would like to setup something that is scaleable and balanced but I >>> don't want to use a bunch of large servers if multiple smaller ones will >>> make the site just as fast. >>> >>> I was considering the following: >>> >>> - Put the forums on nginx instead of apache >>> - using sphinx for search (how does that affect things) >>> - using S3 for profile pictures and such (how does that affect >>> things) >>> - move database off RDS to master and slave setup to save costs >>> - possibly not use amazon load balancers if better to not... >>> >>> Can someone help suggest a setup. What size servers should I do and how >>> many? What do you think about taking the DB off of RDS? etc... >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "scalr-discuss" group. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msg/scalr-discuss/-/H3g6ZPvxFJIJ. >>> To post to this group, send email to [email protected]. >>> To unsubscribe from this group, send email to >>> [email protected]. >>> For more options, visit this group at >>> http://groups.google.com/group/scalr-discuss?hl=en. >>> >> >> >> >> -- >> >> >> Follow us: Twitter <https://twitter.com/#!/scalr> - >> Facebook<https://www.facebook.com/Scalr> >> - Blog <http://blog.scalr.net/> >> >> -- > You received this message because you are subscribed to the Google Groups > "scalr-discuss" group. > To view this discussion on the web visit > https://groups.google.com/d/msg/scalr-discuss/-/T4sIa_QiSxcJ. > To post to this group, send email to [email protected]<javascript:> > . > To unsubscribe from this group, send email to > [email protected] <javascript:>. > For more options, visit this group at > http://groups.google.com/group/scalr-discuss?hl=en. > > -- You received this message because you are subscribed to the Google Groups "scalr-discuss" group. To view this discussion on the web visit https://groups.google.com/d/msg/scalr-discuss/-/JLJA3KM3QiUJ. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/scalr-discuss?hl=en.
