Just saying, if you're doing this it might be a good idea to not have the 'person' closing the thread read as "guest" because to me it looked like a hack
"continuation threads"
- Thread starter tybin
- Start date
- th_trl_thread_readers 0
- Status
Suppliers are really getting screwed big time here because the first post usually is so old it has nothing usefull to a new reader. A lot of time i use the search thread feature which is also now rendered useless. I never had a speed problem so i dont know whats up. Just hope it all can be reversed for everyones sake
Look at the Lavatube thread . From reading the first post a reader would have absolutely no clue what a lavatube is and it was a special and highly anticipated 1 million views that just died. Now you get a thread that starts with broken links.
Look at the Lavatube thread . From reading the first post a reader would have absolutely no clue what a lavatube is and it was a special and highly anticipated 1 million views that just died. Now you get a thread that starts with broken links.
I knew the Lorilland/ Blu purchase was going to be our demise!!! RUN FOR THE HILLS. Bury your PVs in the back yard hide all your juice THE SKY IS FALLING Philip Morris is hovering above my house right now in a helicopter.... HELP
Maybe you're right, maybe not. All I know is that there is considerable debate as to whether long threads cause an issue in
vbulletin. Most reports of very slow load times are with respect to long threads, hence us trying this out.
As to our infrastructure - it is being upgraded, and will happen over the next couple of weeks. Trench, are you a developer? If so, we are always happy to listen to advice.
vbulletin. Most reports of very slow load times are with respect to long threads, hence us trying this out.
As to our infrastructure - it is being upgraded, and will happen over the next couple of weeks. Trench, are you a developer? If so, we are always happy to listen to advice.
Any developer could tell you that a huge thread will not slow down the system. The thread size has nothing to do with anything. A thread is virtual.
Regardless of the post count of a thread, every page query is the same. There's a skip number and a take number. Those two numbers make up the query. The query gets sent to the database, the database returns posts.
The problem lies in either the database or the rendering of the html (ie. too many plugins).
To blame the problem on something that has absolutely no bearing on the rendering of a page is ridiculous.
I read a post by an admin that stated something along the lines of: "More servers/cloud servers will not fix the problem...". And then proceeded to copy paste some words that reeked of "works on my machine". That's a load of you know what.
My guess is that the ECF infrastructure sucks and/or the database is fragmented.
Before you go doing anything else that's pointless, why don't you talk to someone who knows about tuning software? This is a very big and busy forum. I'd put money that it's hitting the limits of the hardware running it.
Again, apologies. The system should not have gone live without our double checking the settings.
One of my threads was split waaay before 5000. It was the 35-54 year old thread.
Yes..I don't think it went as well as they hoped...some were small..must have been a little 'glitch'..! But it's good to see info from what sounds like someone from ECF? I am assuming.....hope it all gets straightened out ....
Ok, now you are someone I'd like to hear more from. Any chance you could PM me details of your server set up etc?
Any long threads that had useful tagging to link to posts of interest also just got rendered worthless as the original threads will just fall off into the abyss.
Just to add something I've mentioned before, you may want to figure out how to stop the forum reloading on every page action. There's no need to reload the previous page every time somebody hits their back button, if anything that'll be adding a huge amount of unnecessary load for the servers.
Last edited:
As part of our programme to improve pageload speed we decided to try all avenues even if in theory there should be no benefit. I completely agree that splitting long threads, in theory, has no possible benefit. This is obvious when you know how it all works. However theory does not always transfer into practice and therefore I am always open to ideas that in theory cannot work - we will still try them. It's just possible that for some reason there might be an effect - even if it all depends, in the end, on a user's PC.
We normally try anything risky on a test server but sometimes for one reason or another we use it live. It might also be a good idea to see if it actually works for anyone before we decide to see if it gets left on. You have to test it live to find that out. Already we have numerous reports that it has magically speeded-up pageload times on long threads - so there you go!
Before the thread splitter goes back on, many of its settings probably need changing. For example the 'Guest' name is not satisfactory and it needs an explanation. I also don't like the way the original post gets to be the first post again in each part - the second post is good enough to be the new second-part starter (with a little adjustment).
We normally try anything risky on a test server but sometimes for one reason or another we use it live. It might also be a good idea to see if it actually works for anyone before we decide to see if it gets left on. You have to test it live to find that out. Already we have numerous reports that it has magically speeded-up pageload times on long threads - so there you go!
Before the thread splitter goes back on, many of its settings probably need changing. For example the 'Guest' name is not satisfactory and it needs an explanation. I also don't like the way the original post gets to be the first post again in each part - the second post is good enough to be the new second-part starter (with a little adjustment).
WHat happens when the old thread that has a wealth of information built up over years gets lost in the pages? Also just thinking about it but i remember reading something about suppliers only allowed to have two threads open. Not sure if thats correct but if so wouldnt that be a issue. A lot of these long threads are like a built up reputation that all goes to waste really. Just thinking of possible issues.
I have not noticed any difference in load times . I havent had a issue in the past though. Maybe its not the server and more peoples obsolete computers and slow internet connection? Is there any actual data showing this really sped things up?
I have not noticed any difference in load times . I havent had a issue in the past though. Maybe its not the server and more peoples obsolete computers and slow internet connection? Is there any actual data showing this really sped things up?
I am just a peon here but I never had any page loading issues. I was a mod on a site where we had 30,00-75,000 people on at any given time and the only thing that ever needed to "upgrade" was storage and bandwidth. So by cutting threads you saved someone with poor internet speed a few seconds of load time, anger quite a few members and suppliers and for what? a few days of faster load times? I will bet in a week if that the people suffering from "slow" load times will have that same problem again. Its called stop being cheap and pay the extra $5-$10 a month and boost your own speed. Just my .02¢
I'm offended -- everyone knows me and calamityjan have the slowest!!!
As part of our programme to improve pageload speed we decided to try all avenues even if in theory there should be no benefit. I completely agree that splitting long threads, in theory, has no possible benefit. This is obvious when you know how it all works. However theory does not always transfer into practice and therefore I am always open to ideas that in theory cannot work - we will still try them. It's just possible that for some reason there might be an effect - even if it all depends, in the end, on a user's PC.
We normally try anything risky on a test server but sometimes for one reason or another we use it live. It might also be a good idea to see if it actually works for anyone before we decide to see if it gets left on. You have to test it live to find that out. Already we have numerous reports that it has magically speeded-up pageload times on long threads - so there you go!
Before the thread splitter goes back on, many of its settings probably need changing. For example the 'Guest' name is not satisfactory and it needs an explanation. I also don't like the way the original post gets to be the first post again in each part - the second post is good enough to be the new second-part starter (with a little adjustment).
Roly, just get rid of the plug-ins like I've been begging you to do for months -- you want speed? You'll get speed then and we all know it.......
You know its just like the government, buddy -- we'll just be told it has worked without any empirical evidence!!!
I feel sorry for the vendors more than anyone else.......
Maybe do an announcement banner a day or two ahead before making changes that are going to obviously be seen/commented on by the users? This was not some kind of emergency patch that absolutely had to happen "right now". Might as well let us know beforehand.
- Status
Similar threads
- Locked
