Support » Alpha/Beta/RC » WordPress MU for 14000 users?

  • Hi,

    I see that Worpress MU creates a bunch of new tables for each and every users.

    wp_[id]_commentmeta, wp_[id]_comments, wp_[id]_links, wp_[id]_options ...

    Is this a scalable option for a medium to large multi user wordpress?
    If I want to migrate my +10K to wordpress MU will my server, database, not moan to much about having +90000 tables?

    Regards

    FFMG

Viewing 10 replies - 1 through 10 (of 10 total)
  • Moderator Ipstenu (Mika Epstein)

    (@ipstenu)

    🏳️‍🌈 Halfelf Rogue & Plugin Review Team Rep

    That’s not for each user, that’s for each blog on your MU install.

    If you want to support 14000 BLOGS, you’ll need a beefy server. Turn off the ability to have users create blogs when they make their accounts, and that won’t be an issue 🙂

    Hum… I already have +14000 blogs, (with my current software I have one user/one blog).

    I want users to create a blog when they sign-up

    FFMG

    Moderator Ipstenu (Mika Epstein)

    (@ipstenu)

    🏳️‍🌈 Halfelf Rogue & Plugin Review Team Rep

    Then you are going to need a powerful server with a lot of memory. Yes, it’s scalable (wordpress.com is running a version of MU IIRC) but it’ll need a hefty server.

    (PS – Don’t sign your posts. We can see who you are by looking to the left of your post 😉 )

    Then you are going to need a powerful server with a lot of memory. Yes, it’s scalable (wordpress.com is running a version of MU IIRC) but it’ll need a hefty server.

    I am a little worried about that approach, throwing hardware at the problem might not be the best solution.
    I simply don’t know enough about MySQL to know if that is a better solution rather than the same tables with an extra Unique user ID row.

    Is it really practical to have +90000 tables on a MySQL database, (in my case)?
    Is that not a problem in itself?

    Do you, (or someone), know if there has been some kind of discussion about using that approach?
    I just want to re-assure myself that I will not kill my servers because of a bad design to start with.

    (PS – Don’t sign your posts. We can see who you are by looking to the left of your post 😉 )

    It feels a bit rude not to, I don’t know of any forum that advises _against_ signing your posts.

    FFMG

    Moderator Ipstenu (Mika Epstein)

    (@ipstenu)

    🏳️‍🌈 Halfelf Rogue & Plugin Review Team Rep

    (PS – Don’t sign your posts. We can see who you are by looking to the left of your post 😉 )

    It feels a bit rude not to, I don’t know of any forum that advises _against_ signing your posts.

    Now you do: http://wordpress.org/support/topic/374352

    E: Signatures
    Signatures on all posts will be removed. Signatures create clutter and distract from the information and help we are trying to provide.

    Topic on hand.

    I’m not saying that throwing hardware at the problem is the solution, but I am saying this: If you’re going to run a site where 14000 people can login and post blog stuff, then YOU NEED A GOOD SERVER. If you said ’14’ people, I wouldn’t have mentioned. At 140 I’d probably suggest a VPS. And 1400? Dedicated server time.

    This has nothing to do with WordPress scalability. It’s just good, server admin, sense.

    I already have a dedicated server with a fair amount of memory, (admittedly it is running a bit high at the moment).
    but I would hate to migrate 14000 blogs only to find that I need a much beefier machine than the one I am using already.

    And given that I am planning to run a beta for 6 months or so I will effectively need to support 28000 blogs.

    Ideally there would be a way of working out what would be needed _ahead of time_.

    … It’s just good, server admin, sense.

    I understand what you are trying to say, but a good project manager, a good developer and a good server admin would want a ball park figure of how much hardware is needed ahead of time rather than simply trusting that mu.wordpress is running efficiently.
    That, to me would be what a good server admin does.

    Err, yeah, but if you have a specific target in mind then generally you’d check it out on a testing/scalability platform of your own. It’s not likely that others will have reproduced your own specific setup and use cases.

    More and more MU discussion on this forum, not? Better direct MU users here?
    http://mu.wordpress.org/forums/

    More and more MU discussion on this forum, not? Better direct MU users here?
    http://mu.wordpress.org/forums/

    I agree with you, but now that the line is a bit, (more), blurred with version 3 of wp, maybe a forum should be created here for beta MU version 3.

    But, as my particular question is not really Beta related, maybe I should start a new topic there or it should be moved.

    Andrea Rennick

    (@andrea_r)

    Customer Care at Copyblogger Media and Studiopress

    It’s not beta related, as it’s pretty much a database scaling questions, which we’ve gone over quite in depth (repeatedly) over in the WPMU forums.

    Remember, wordpress.COM has millions of users & blogs. they’re using the same software.

    Is it really practical to have +90000 tables on a MySQL database, (in my case)?
    Is that not a problem in itself?

    You don’t use one database at that size. You replicate or split them. There’s three plugins that currently exist for this, including the one wp.com uses.
    http://wordpress.org/extend/plugins/hyperdb/
    http://wordpress.org/extend/plugins/shardb/

    And the other issues is you’ve said you have a “good dedicated server” but you didn’t give any specs at all. That big of an install, I’d probably start with two boxes, one just for mysql.

    again, though, we’ve gone over this in the MU forums plenty of times.
    http://www.google.com/search?sitesearch=mu.wordpress.org&q=scaling

    maybe a forum should be created here for beta MU version 3.

    there’s a forum for the beta in general. MU will cease to exist, there’s not a “beta MU version 3”. there’s just beta 3.0 with networks enabled.

    Until the current MU forums roll over here, might as well go over there.

Viewing 10 replies - 1 through 10 (of 10 total)
  • The topic ‘WordPress MU for 14000 users?’ is closed to new replies.