Transparent Moderation Log & Site Topics - Part II

Part I was getting too big and choking server. Continued. Original thread is here. Moderators: please post a brief explanation when a thread move, lock, post deletion or banning is performed. Members may check here to see what happened. Hopefully, this will cut down on the "what happened to x" clutter. This is not a thread to call out members, to say "well if he got banned why didn't X,", etc. Concerns about members or threads should be expressed in private message to me or to another appropriate moderator (for the subforum in question). Thanks.

Part of the blame for the recent lag goes to humongous threads, some with over 10,000 replies, which also tend to be ones that draw a disproportionate share of traffic. Therefore, I have cut some of the larger threads off into Part II/III etc. and will continue to do so. Normally I would move the last few pages into the new one, but on this one I didn't see much point.

According to the guy I talked to who seemed to know his stuff, vBulletin and/or MySQL don't handle the huge threads very well, especially when there are concurrent requests for them (i.e. a ton of people trying to read/post on a thread with 10,000 replies). Some forums actually employ a hack that automatically splits a thread after X replies - after the coming upgrades/optimization, we will see if we need to do something like that. Edit: you changed your question in the middle of me answering. I don't think the short new threads will draw less traffic than the long ones did, but they are shorter, and hence the server is more capable of dealing with it. Think of it as if you had a huge book with 10000 pages and had 1000 people running up asking you to tear out page number X. If your book was 1000 pages instead, it would obviously take you less time to find their page and give it to them.

The hard thing is that every time we 'fix' or upgrade the server so there is little lag for a few days in a row, we get a huge spike in traffic that doesn't go away, thus putting more load on it at all times thereafter, and increasing the chance of another problem or just overloading it. So this next upgrade, we are going to focus on way overdoing it to future-proof. I'll be very interested to see how our traffic numbers look when the server can put up with anything that can be thrown at it. Based on the current trend, it could get ridiculous, but I'm sure there will be an equilibrium point - we just haven't found it yet.

As always, contributions to the cause (server upgrades, hosting fees, my eventual liver transplant) can be made to j2@styleforum.net (for normal paypal balances) or j@styleforum.net for anything more complicated (it takes more fees, is all). Thanks for thinking of us.