"continuation threads"
- Thread starter tybin
- Start date
- th_trl_thread_readers 0
- Status
and for what it's worth - my page load times on those short threads are atrocious this am ......
You'll have to read the entire thread to find that answer.
Post from Smokey Joe
Hey all, this was an update that took place last night. I'm not sure what happened, but it looks as if it was set to 'live' before it had been configured properly.
Our intention is to have the splitter split the threads when they reach 5000 posts. When configured properly it will credit the original poster, and have links to the old thread from the new thread, and new threads from the old.
The reasoning here is that there is some evidence that very large threads are causing database issues and adding to the overall server load - causing slowdowns at busy periods. This is the first of 3 changes we are going to be making over the next 3 weeks to improve the load speed for everyone.
I'm sorry this happened before an announcement was made.
Last edited:
The lack of consistency on when a thread is closed with new one started is crazy. Was it supposed to be 5000 posts? 500? (500 is way too small IMO, but I'm seeing threads in the 1 - 2K post zone got split).
From some quick searching, it seems like vbulletin may (only may) indeed have resource issues with large threads. Assuming the cap was actually implemented at 5K posts, if it overall improves responsiveness while impacting a relatively small number of the threads (however much some love them), isn't that a reasonable thing to do? I'm pretty active, including some threads in the 10s of thousands of posts, but honestly if new threads got started every 5K it wouldn't cause huge difficulty.
To be effective any cap really should apply across the board -- no "cherry picking". Vendors, social groups, all should be affected equally.
I don't know how many 5K threads there are vs. 10K, or if the 10K-class threads are still generating a large number of hits (people may be checking out the monster threads even if not directly adding posts). Maybe start with a cap at 10K, see if it improves life over a while, and if not make a try at 5K.
From some quick searching, it seems like vbulletin may (only may) indeed have resource issues with large threads. Assuming the cap was actually implemented at 5K posts, if it overall improves responsiveness while impacting a relatively small number of the threads (however much some love them), isn't that a reasonable thing to do? I'm pretty active, including some threads in the 10s of thousands of posts, but honestly if new threads got started every 5K it wouldn't cause huge difficulty.
To be effective any cap really should apply across the board -- no "cherry picking". Vendors, social groups, all should be affected equally.
I don't know how many 5K threads there are vs. 10K, or if the 10K-class threads are still generating a large number of hits (people may be checking out the monster threads even if not directly adding posts). Maybe start with a cap at 10K, see if it improves life over a while, and if not make a try at 5K.
Mark - it was set at 500 last night. It's off now and we're discussing what we are going to do going forward.
Thanks for your suggestions...
SJ
Thanks for your suggestions...
SJ
I'm betting it won't be too long before this thread is 'split' also!
Last edited:
I have to say I was a bit puzzled when I saw this???
Anyway, if Smokey Joe is on it, and Roly is posting at 7 am , I'm pretty confident that some of the best Web/Forum guys in the world are working on it...Maybe I'm a little naive but, I've been a member here for a few years and even under hostile attack the forum has always worked pretty well and never really stayed broke for long.
Anyways good luck with that... I'm glad I don't have your jobs!!!!
Anyway, if Smokey Joe is on it, and Roly is posting at 7 am , I'm pretty confident that some of the best Web/Forum guys in the world are working on it...Maybe I'm a little naive but, I've been a member here for a few years and even under hostile attack the forum has always worked pretty well and never really stayed broke for long.
Anyways good luck with that... I'm glad I don't have your jobs!!!!
Hey Stranger!!! Is that the only way we get to hear from you these days? We gotta break somethin? Good to see ya around.
Drew
I am just curious about something.
It seems like some of the bigger threads have been re-opened. I am just curious why?
The main reason I ask is that I do not necessarily want to add the new "part 2's" or delete the old ones yet if they are going to be re-opened.
You might not even have the final answer yet. I was just curious. thanks in advance!
It seems like some of the bigger threads have been re-opened. I am just curious why?
The main reason I ask is that I do not necessarily want to add the new "part 2's" or delete the old ones yet if they are going to be re-opened.
You might not even have the final answer yet. I was just curious. thanks in advance!
Hmm...
The thread splitter will most likely go back on, because some people said they had an improvement - and that is all we are trying to do. Also, no one has come in and said, "This has wrecked my threads/forum/life". We will wait, though, to see if there are any negative issues that have to be addressed. We need to look at several split threads in different locations and see what the result is.
However there are all sorts of issues with it that we can see, now it's been trialled:
- Exactly what figure to set it at. If indeed it works at all (about which there will always be arguments no matter what happens), setting it too high would be a waste of time, and setting it too low will cause too much disruption. It looks as if 2,000 might be a good compromise number: 200 pages of default thread size. No one will be inconvenienced if a thread is split into 200-page parts.
- The name of the 'splitter admin' or whoever it is the makes the first post in the new part is 'Guest', and that's not too good
It needs another ID.
- The first post in the subsequent parts is currently the first post in the thread, and I don't think this is a great idea. It could be left off, and just have the: 'This is Part 2, here is the link to Part 1' post - this would be enough. This will require a re-code but it shouldn't be too hard to fix this.
- There is at least one issue for Suppliers: Part 2 onwards does not now 'belong' to the OP, so if it's a Supplier post in a dedicated forum, they can't close/open the post. Not sure how much impact this will have.
There is some fixing to do before it goes back on for trial #2. Then the fun will start.
The thread splitter will most likely go back on, because some people said they had an improvement - and that is all we are trying to do. Also, no one has come in and said, "This has wrecked my threads/forum/life". We will wait, though, to see if there are any negative issues that have to be addressed. We need to look at several split threads in different locations and see what the result is.
However there are all sorts of issues with it that we can see, now it's been trialled:
- Exactly what figure to set it at. If indeed it works at all (about which there will always be arguments no matter what happens), setting it too high would be a waste of time, and setting it too low will cause too much disruption. It looks as if 2,000 might be a good compromise number: 200 pages of default thread size. No one will be inconvenienced if a thread is split into 200-page parts.
- The name of the 'splitter admin' or whoever it is the makes the first post in the new part is 'Guest', and that's not too good
It needs another ID.- The first post in the subsequent parts is currently the first post in the thread, and I don't think this is a great idea. It could be left off, and just have the: 'This is Part 2, here is the link to Part 1' post - this would be enough. This will require a re-code but it shouldn't be too hard to fix this.
- There is at least one issue for Suppliers: Part 2 onwards does not now 'belong' to the OP, so if it's a Supplier post in a dedicated forum, they can't close/open the post. Not sure how much impact this will have.
There is some fixing to do before it goes back on for trial #2. Then the fun will start.
Last edited:
Hey Drew - I'm quietly lurking in the background right now, but gearing up to spending a lot more time here. Good to see you too!
Actually we don't know. Because of recent interest in pageload speed issues, we decided to try various things to see if there is any benefit for some people, without impacting the majority.
This trial has been extremely useful because the main thing we learnt is that some people - the people complaining about long thread opening times - say it's an improvement. If they don't, well, it's no use.
However we really don't know how this will pan out - what the part size will be finally agreed as, for example (100 pages, 200 pages, 500 pages or what). So I apologize about the bookmark situation but until we have settled a couple of issues, we don't know the answer. *Page* bookmarks in long threads aren't going to be much use for a couple of days
Mind you it could be easier to bookmark a post at the point you want to get to - that won't change. If it does we have real problems
Last edited:
rolygate,
just my 2c but I have not necessarily noticed anything moving faster. I want you to know I am not complaining, I just have not noticed a difference. I think I read that you guys wanted feedback on that.
I wonder if now having double of everything could cause problems as well? I am certainly no expert in this field so take what I am saying with a grain of salt
the main issue now is the number of subscriptions that we will all have. and how to keep those straight.
thanks for the response.
just my 2c but I have not necessarily noticed anything moving faster. I want you to know I am not complaining, I just have not noticed a difference. I think I read that you guys wanted feedback on that.
I wonder if now having double of everything could cause problems as well? I am certainly no expert in this field so take what I am saying with a grain of salt
the main issue now is the number of subscriptions that we will all have. and how to keep those straight.
thanks for the response.
Hi There,
Might as well post here...
OK, so here's how we fix this---Give us (the Suppliers) back the option to edit the old lead posts, and the option to close the spurious Part2 threads...
The way it is now, every supplier leads off with old product and most started with low ball prices just to get a foot in the door.
This creates ill will in that these sale/grand opening prices are no longer available...
I'd be happy to start a updated thread, but only an updated thread with just a link to the old...and its lead post.
I cannot as it is edit the lead post or close my own thread.
This is the problem---it's forced confusion, and we Suppliers take the hit (not good biz)
Seems like a simple fix---and sure I hate the fact I lose all my street cred by being seen as a new Supplier..
I've build a reputation here over years and do not feel very good about starting over...
But I will, just give me back some control, so as to not be saddled with the insult of saying no to every viewer that expects to pay day one prices...
That's bad mojo, and It's dead wrong to inflict that on me...
I'm a dedicated A+ Supplier and deserve better treatment
Thanks
fonosmoak
Creator of the very first commercially available High End exotic wood PV
i.e The Phidias, a totally original design
Thanks
fono
Might as well post here...
OK, so here's how we fix this---Give us (the Suppliers) back the option to edit the old lead posts, and the option to close the spurious Part2 threads...
The way it is now, every supplier leads off with old product and most started with low ball prices just to get a foot in the door.
This creates ill will in that these sale/grand opening prices are no longer available...
I'd be happy to start a updated thread, but only an updated thread with just a link to the old...and its lead post.
I cannot as it is edit the lead post or close my own thread.
This is the problem---it's forced confusion, and we Suppliers take the hit (not good biz)
Seems like a simple fix---and sure I hate the fact I lose all my street cred by being seen as a new Supplier..
I've build a reputation here over years and do not feel very good about starting over...
But I will, just give me back some control, so as to not be saddled with the insult of saying no to every viewer that expects to pay day one prices...
That's bad mojo, and It's dead wrong to inflict that on me...
I'm a dedicated A+ Supplier and deserve better treatment
Thanks
fonosmoak
Creator of the very first commercially available High End exotic wood PV
i.e The Phidias, a totally original design
Thanks
fono
Last edited:
Roly, just a thought, but you really might want to start higher, up around 5K, let it run a handful of days, only then drop it one more notch if need be.
Assuming what I have read is true, that vbulletin likes to queue everything for a given thread when it is opened, logically starting with the largest active threads might have a lot of bang for the buck while minimizing impact overall. There are a relatively small number of huge traffic threads that may well be using more resources.
The people who daily are viewing or commenting in the really big threads may be having much more impact than the threads down in the 2 - 5K zone. I'll pick on myself -- I frequently am in a chit chat thread with about 45K posts, will refresh/read multiple times throughout a day even if I'm not commenting. If VB is doing anything relative to the whole record set when this occurs, it would be similar to 9 people adding traffic to a 5K post thread. By their nature the really large > 5K post threads are also the ones that have built up more community "mindshare".
Assuming what I have read is true, that vbulletin likes to queue everything for a given thread when it is opened, logically starting with the largest active threads might have a lot of bang for the buck while minimizing impact overall. There are a relatively small number of huge traffic threads that may well be using more resources.
The people who daily are viewing or commenting in the really big threads may be having much more impact than the threads down in the 2 - 5K zone. I'll pick on myself -- I frequently am in a chit chat thread with about 45K posts, will refresh/read multiple times throughout a day even if I'm not commenting. If VB is doing anything relative to the whole record set when this occurs, it would be similar to 9 people adding traffic to a 5K post thread. By their nature the really large > 5K post threads are also the ones that have built up more community "mindshare".
Roly, and Smokey: I understand your desire to speed things up, but honestly there are allot of vets posting in THIS thread saying that this may not be the best alternative. an invaluable amount of information lies in those threads. I think giving newbies that kind of research enables them to see the questions that were originally asked and assists them in making an informed decision before they purchase. I hope you will reconsider. Thanks
Yes Mark.
These are things we need to try and figure out.
Also it now looks as if we may have to exempt Supplier threads from this (as we can). We don't want to hurt people in any way.
Everything has a cost/benefit balance, at first this looked as if it would help - but there may be too many negatives. I personally am not convinced that this can provide any speed increase, from the technical point of view, but I need to repeat that with the recent strong interest in pageload times - and especially the perception that it is not a local issue - we have to try various things to see if any improvement results.
Also, even when theory says something is not possible, sometimes it does work...
These are things we need to try and figure out.
Also it now looks as if we may have to exempt Supplier threads from this (as we can). We don't want to hurt people in any way.
Everything has a cost/benefit balance, at first this looked as if it would help - but there may be too many negatives. I personally am not convinced that this can provide any speed increase, from the technical point of view, but I need to repeat that with the recent strong interest in pageload times - and especially the perception that it is not a local issue - we have to try various things to see if any improvement results.
Also, even when theory says something is not possible, sometimes it does work...
Last edited:
- Status
Similar threads
- Locked
