It is NOT down to internet speed

[eFestivals]

Why would the MacBook via the same connection behave so differently to how it did on Wednesday?

The answer, of course, is that it wasn't the MacBook which was behaving differently, it was what the MacBook was responding to (or trying to respond to) which was behaving differently - the See servers that it was trying to communicate with.

This is the part that those who like to think they have some sort of 'special' method never factor into their thinking, just as they never factor in the fact that the Macbook is just a computer like a PC is computer and a phone is a computer, and that an ISP is just an ISP whether via a landline or a mobile network. In all cases they're all using the same common-standards technology to create and send and receive and act on what are merely just common data requests.

[tweed1981]

I personally do not think I have any sort of special method in the slightest, I just find it odd that some people report pretty much just a blank screen or variations of not a lot happening, and others like myself have instances of everything refreshing quickly and working perfectly.

One method worked like this on Wednesday and the exact same method didn't on Sunday, whereas another method did. I'm not suggesting some massive conspiracy rather that some requests are dealt with differently to manage the load.

[scrippit]

I've always believed it to be entirely down to luck. Until you mentioned 'wearing special pants'.

I was indeed 'wearing special pants'. Thus all the evidence points to the fact that the wearing of special pants improves your chances of getting to the holding page. I love science.

[eFestivals]

I just find it odd that some people report pretty much just a blank screen or variations of not a lot happening, and others like myself have instances of everything refreshing quickly and working perfectly.

As I pointed out, it's what the other end of things (See's servers) are doing that causes any browser to act as it does.

For those getting a blank screen, they're not getting a satisfactory response from See's servers for whatever reason.

My guess (based on what can happen with efests on high-load days) is that See's servers have reached their capacity serving other people - and it seems to be the case that once your browser has been 'locked out' of their system with a white screen it remains 'locked out' until something (either in the browser, or at See's servers) times-out so that it can start again afresh.

The best thing, I think, to do if you get that white screen is move to another machine or fire up a different browser on the same machine - and that will then be the same as that original browser starting afresh.

This seems to be borne out by the number of people who say they constantly get a blank screen but who then get straight thru when they try another browser or another machine.

That's the only thing which makes any sense to me from a technical point of view - tho while i'm reasonably confident in my knowledge, I wouldn't claim to be an expert.

[billum]

I've always believed it to be entirely down to luck. Until you mentioned 'wearing special pants'.

I was indeed 'wearing special pants'. Thus all the evidence points to the fact that the wearing of special pants improves your chances of getting to the holding page. I love science.

In the light of all the other evidence I was starting to be convinced by this pants theory, until I realised I actually got through twice this year going commando.... 8-0

(tmi I know)

[amfy]

This is obviously what went wrong for me this year. I wasn't wearing any pants.

[scrippit]

In the light of all the other evidence I was starting to be convinced by this pants theory, until I realised I actually got through twice this year going commando.... 8-0

(tmi I know)

Darn evidence based analysis!

[tweed1981]

This is what I'm saying about being lucky twice, you may initially not get the correct response from the see servers, so you may then get locked out. You may be sitting there with a blank screen pressing f5 with no real hope of getting anywhere... unless you shut it down and try something else. That is all very well if you know to do that,but if not then luck is very much against you compared to somebody who got the right response first of all, or closed down and tried something else and then starts to get the holding screen.

[jamesrfisher]

I had 4 browsers open, each with only one tab. I had the holding page on one and was on the ticket page by 9.03 with payment made by 9.05.

Why was this? I put it down to having my lucky EFests t-shirt on

[Sawdusty surfer]

I had 4 browsers open, each with only one tab. I had the holding page on one and was on the ticket page by 9.03 with payment made by 9.05.

Why was this? I put it down to having my lucky EFests t-shirt on

Well, if wearing the eFests t-shirt worked, clearly you must have been wearing it like this.

[siblin]

It doesnt make any difference at all. TCP is a democratic protocol first in first out. We had connections from 4 different ISPs and 1 Gb company connection. Got tix on a samsung tablet on the slowest ISP

FIFO - therefore latency has an effect.

But the effect of the load balancers is far more of a concern. likely posters that claim moving to a different connection helped was more to do with something called "sticky sessions" which i wont go into as its very boring.

[Sack truck]

I'd imagine a few folk have had sticky sessions on the web they'd rather not talk about

[eFestivals]

FIFO - therefore latency has an effect.

But the effect of the load balancers is far more of a concern. likely posters that claim moving to a different connection helped was more to do with something called "sticky sessions" which i wont go into as its very boring.

I've just gone and read up a little on that, and it seems to be saying that my 'informed guess' (that I've posted in this thread or others) for why some people were getting a white/blank screen is about right - that those people aren't getting allocated a server session for whatever reason (server overload being the obvious one).

Would you agree with that?

[Pinhead]

I kept getting a white screen for each refresh on some browser sessions, rather than the holding page. I put this down to the browser tbh rather than it being a See issue (other than at times it was too busy to even serve the holding page to clients). Flushed the cache on the browser, closed it, opened it again and issue seemed to go away and I got the holding page again.

[eFestivals]

I kept getting a white screen for each refresh on some browser sessions, rather than the holding page. I put this down to the browser tbh rather than it being a See issue (other than at times it was too busy to even serve the holding page to clients). Flushed the cache on the browser, closed it, opened it again and issue seemed to go away and I got the holding page again.

How could it be a browser issue? ... Surely if browsers were the cause, they'd regularly do that on other pages too?

The browser responds to the data that comes back (or doesn't) from the data requests it sends out, so I can't see how it's a browser thing. It's much more likely to be about the data it gets back to my mind.

Clearing the cache and restarting will clear out anything malformed that's come back from a data request if malformed data is the problem, and will clear any session related stuff which might be causing it to stay as a white screen on continual requests if it's server session related stuff - so a browser restart will be starting afresh at both ends, which will give better luck if the servers are then able to properly serve data back at that later time.

This is why, so i'm thinking, so many people end up believing that mobile devices have an advantage, cos for many people their phone or tablet will be their alternative device to try (which is starting afresh, just as with a clean and restarted browser).

Edited October 9, 2014 by eFestivals

[natdan73]

Someone on here once said that Internet Explorer is not good at handling very heavy traffic - whether that is true I do not know

[eFestivals]

Someone on here once said that Internet Explorer is not good at handling very heavy traffic - whether that is true I do not know

IE is often slower to display the HTML that it gets back from a server than other browsers, because of how it processes that HTML. But this difference is pretty minimal in most cases, to the point that it can basically be ignored.

There's nothing 'heavy traffic' about the Glastonbury sales process from what happens at the browser end of things as far as different browsers are concerned.

[Pinhead]

How could it be a browser issue? ... Surely if browsers were the cause, they'd regularly do that on other pages too?

The browser responds to the data that comes back (or doesn't) from the data requests it sends out, so I can't see how it's a browser thing. It's much more likely to be about the data it gets back to my mind.

Clearing the cache and restarting will clear out anything malformed that's come back from a data request if malformed data is the problem, and will clear any session related stuff which might be causing it to stay as a white screen on continual requests if it's server session related stuff - so a browser restart will be starting afresh at both ends, which will give better luck if the servers are then able to properly serve data back at that later time.

This is why, so i'm thinking, so many people end up believing that mobile devices have an advantage, cos for many people their phone or tablet will be their alternative device to try (which is starting afresh, just as with a clean and restarted browser).

Yes, I broadly agree with this. I too, cannot see how a browser would get hung up like this either, though I offer the experience only as anecdotal evidence rather than proof that there is a browser design issue here. IE seemed worse for this than other browsers (tried Firefox, Chrome and Opera). Of these, Opera surprisingly gave the impression of being the most stable.

[siblin]

I've just gone and read up a little on that, and it seems to be saying that my 'informed guess' (that I've posted in this thread or others) for why some people were getting a white/blank screen is about right - that those people aren't getting allocated a server session for whatever reason (server overload being the obvious one).

Would you agree with that?

I would yep.

Okay, I'm sure you probably know some of this Neil (and it is very boring), but I'll go into it anyway for the 3 people in the world that might be interested . Load balancers usually have some algorithm behind how they balance sessions. Simple ones (commonly weighted round robin, least sessions) or more complicated ones, (weighted least connections for example) each one has limitations that can be quite severe under over-load conditions - a server that has crashed can often have the least number of active connections as its just bouncing them straight back. There are all manner of healthchecks that go on to try and mitigate this. it's most likely that see are using some form of WRR as their algorithm.

WRR algorithm (as with most LB algorithms) basically hands each session that comes in around a pool of backend servers. Problem is in a system such as one buying tickets, you want to track session ID's, and cache database queries, which means randomly jumping between servers during each HTTP request could cause issues. To combat this LB's have a feature called sticky sessions which track on the loadbalancer (or 3 in the case of see) based on the session criteria (usually source/destination IP and TCP port) what backend server to direct your traffic toward.

Problem is If that server is overloaded it doesn't matter how many times you try and refresh you'll be pointed at the same back-end. Fire up a different device, use 3G instead of your broadband etc and you'll be handled differently, which I think accounts for the "I fired up my ipad and it worked instantly!" anecdotes.

[Pinhead]

Now this is the best explanation I've heard in a while, that does seem to be supported both by some anecdotal experience on the day and by the known nature of the architecture of the system See use for load balancing.

[LondonTom]

Yes, I broadly agree with this. I too, cannot see how a browser would get hung up like this either, though I offer the experience only as anecdotal evidence rather than proof that there is a browser design issue here. IE seemed worse for this than other browsers (tried Firefox, Chrome and Opera). Of these, Opera surprisingly gave the impression of being the most stable.

Eliminates the browswer being the problem (and Neil's and Siblins load balancing stuff) for me! As Opera didn't get no where near for me, just a white page.

While IE, Firefox and Chrome all got to at least the queue page.

[eFestivals]

I would yep.

Thanks for the explanation, and it's good to hear that my ideas seem to be in the right ballpark.

[JGP]

I have said this before but it made perfect sense to me when another poster put it this way......

Its like trying to find a parking space by driving round and round the block. Whether you are in a souped up formula one racing car, or crawling round in a clapped out Robin Reliant, you will get a space if you are passing when one becomes free. The faster car can speed up its chance of missing a space, as much as it speeds up its chances of finding one.

what has increased my chances? In 2005 I missed out. In 2007 I was lucky, in 2008 I started helping other random people, from here, from work, anyone who said to me 'So how do you get Glastonbury tickets?', anyone who wailed on here or fb that 'they weren't getting a sniff of a booking page' - I helped if I could. - & I carried on being lucky, until this year when my luck ran out.

...

.but that was when it all came back & no less than 3 of the people I'd got a booking page for in the past, got one for me.

yes networks do improve your chances. I am not incredibly popular. I do not have a 'Glastonbury crowd' - but over time, I have increased my chances of luck.

I honestly think that's he only thing that you can do - work on increasing your network. Nothing to do with computers.

Using the car park thing again, if you've got a quicker car you can drive past more spaces and therefore have more opportunities to find a space?

[Jack Nottm]

So come the re sale or next year say. If i get a white screen on say firefox, its likely i have been given a sticky session. In order to get round this i would.... clear the cache, close the browser and open it again? move to a different browser totally?

For instance this year, every device, 1 desktop and 3 laptops white screened dead on 9am. Phone on 3g was getting refresh page but was taking a while. After 5 minutes i saw that none of the white pages were going anywhere and loaded up a tor session which got the 20second page and then eventually got through by manually refreshing.

Also, just had a thought. Would it ever be worth waiting till say 2 mins past 9 to try and get on the page since a load of people will have been stickied at 9 and so lots of people will no longer be able to hit the server...

Edited October 11, 2014 by Jack Nottm

[eFestivals]

So come the re sale or next year say. If i get a white screen on say firefox, its likely i have been given a sticky session. In order to get round this i would.... clear the cache, close the browser and open it again? move to a different browser totally?

either of those should do it, tho it would be quicker to fire up a different browser (as long as it was already installed).

Also, just had a thought. Would it ever be worth waiting till say 2 mins past 9 to try and get on the page since a load of people will have been stickied at 9 and so lots of people will no longer be able to hit the server...

that's one of those things where it might give you a better chance but might also do the opposite. There's no way of knowing from a user's point of view which way it is.