Vanilla Performance

ug

So I've found vanilla.. :) Has anyone got any performance numbers for large boards? For example my current board has around 81,000 registered users with 700-1100 online for most of the day. Can vanilla handle something like this?

Mark

Before I'd even think about how Vanilla would perform, I'd think long and hard about how a community of 700 - 1100 regular daily users would react to a forum like Vanilla. Vanilla is very much a "new" forum, and it will take a very savvy audience to actively jump on board. It has been my experience in recent past that when traditional-style forum users are presented with Vanilla, they do not react well - especially if those users are not techie/power users. Most people have an idea in their mind of what a forum is and how it works - and rather than greet a new methodology with open arms, they will retaliate with vigour. Don't get me wrong, I'd love for an audience that large to get a taste of Vanilla, but I also don't want you to have a lot of grief in the process. Vanilla will take some time to catch on, and I believe it will be with smaller communities that it makes it's largest strides. I'd be willing to bet, for example, that you have a lot of different categories (and probably even sub-categories) in your current forum. If so, the "filtered discussions" method of Vanilla might make it harder for people to keep track of what's going on in a community that large. Just some things to consider... As far as stress testing goes, I've done bottleneck testing locally and done my best to remove any that I come across. I have not run any serious stress testing on Vanilla at all. Perhaps the best move for you would be to set up a test version of Vanilla and let some of your regular members try it out. You will both get a feel for how they will like it (or not) and you will also see how that many users impacts performance.

folletto

Could I make a note?

I think that any non-forum user would find Vanilla 10+ better than any vB or phpBB powered forum... so, I think that the most problematic users will be the "core" forum users (vB or phpBB addicted) that will found something "less" than they have used...

Those kind of users usually forgot what a forum is about: communicating, talking. So they think that a forum "with signatures, with big avatars and many buttons" are better than others.

Vanilla is *great* because it's powerful, flexible, clear and simple. If it could handle as well a huge amount of users well... I think that it could be the best forum engine ever.

So...

If someone could test Vanilla under stress... it will be great.


~


Sidenote: I bumped a bunch of friends (about 35 users) from a vB forum to a Vanilla forum, with a Textile extension I've written for it. They adapted in less than one day, and in 3 days they managed perfectly Textile.

jesusphreak

Mark and folletto hit the nail on the head. Most designer/programmer/pragmatic types see the value in Vanilla and love it. My forum of gamers, however, hated it. I ended having to go back to a vB-style board, even though it was against my own feelings. I gotta say I appreciate Mark's openness instead of being like many other people who just madly hype their own product, blindly ignoring all issues with it. Actually, if Vanilla's only issues are the fact that its different and therefore certain people with a bias won't use it, I'd say it is doing pretty good.

lech

I could have sworn that a stress test was done under an older core by someone except we never really had definate results from that test. But regardless of that, I think switching that many users from one system to another without any type of transition, it would most certainly cause a decline or distaste among your users. I would say, if you were to test the waters, run the boards side by side in a comparison test and take a good crash test over the course of a month to see how well things favor for your users. Some will love it, some will hate it, but regardless you'll have a good way to gague how they react if you give them an option before making a full change over. I would love to see some stress stats based upon that test to see how well the forum can handle that kind of abuse. One sure-fire method to ease the transition would be to style it along the current theme of whatever board you're using now. It would be like dumping a dedicated IE users into an IE themed version of firefox, at first, they would hardly notice the difference :)

ug

Wow thanks for all the great comments. Moving to vanilla would not be an overnight deal by any means but setting up a "test" board and pointing some users towards it to get feedback would be the first step we take. While most of our users are used to the normal forum "look & feel" I think a large user base would like the fact we could possibly switch to an open source platform. The extension interface that vanilla provides could allow us to extend vanilla in anyway we saw fit and have it be a community effort.. I did some searching around has anyone written a vb convert utility yet? I'm curious to see how tons of data in vanilla would impact it..

Minisweeper

If you can find a vb->phpBB converter you can use my phpBB->vanilla migration tool. Beware though that its by no means flawless! It should atleast successfully transfer all your data (give me a shout otherwise) so you can see how vanilla copes.

One thing i will add to this, is now i've been using vanilla for just under a year (and o8 before that, so i guess my forum use is fairly defined) i'm starting to really hate using different software. I just joined the WSCC forum which is running iB and it's just so irritating having so much junk everywhere!

ug

minisweeper: thanks I'll give that a go do you happen to have a link to your phpBB->vanilla migration tool? I would have to agree, the bloat in normal forum software drives me insane. I find most of it is disabled on my forum anyways..

Minisweeper

This should do the trick.
Since you seem fairly technically minded if you'd like to tidy it up you're welcome ;p

Mark

One more note on performance: Vanilla 1 will run faster than 0.9.x

Bergamot

Why?

Mark

A lot of improvements have been made to the core and I've done a lot of bottleneck testing, removing some major hinderances that existed in older versions - not to mention a great deal of sql query analyzing and speed improvements on that end as well.

ug

Very cool Mark.. Can't wait to see the final 1.0 release..

Bergamot

Cool

lech

Wewt!