stevegrant

Probably because he's the one online when the (countless) reports about your posts flood in.

Us and Colchester are the only two teams not to have lost a game when scoring first. We've won 12 and drawn 3 games (Orient at home, Yeovil and Peterborough away) when doing so. Surely that's a contradiction. Either we remember we're in League One, and therefore accept that we have no divine right to be "walking the league" regardless of resources and adjust our expectations accordingly, or we act the billy big ******** and then whinge that fans of Brighton and Bournemouth don't like us for some unknown reason.

Taking out any emotional attachment to any particular result, particularly thinking about Tuesday's defeat at Walsall, and ignoring the actual "performances" in terms of fluidity, cohesion, consistency, etc, and solely looking at the end results and the points accrued, I'm confused as to why so many people are so concerned about our current situation. Since the new year, we have picked up 24 points from 12 games, an average of 2 points per game - comfortably promotion form, in normal circumstances. The only problem - and it's something almost entirely out of our control - is the form of other teams. Brighton have averaged 2.33 points per game in the same period (albeit only 4 points more in total) and Bournemouth have averaged 2.21ppg. Leeds went up last season with 86 points, but based on form since the new year (at least 12 games), there are SEVEN teams who would end with more than that over an entire season: Brighton 2.33 (28 from 12) Bournemouth 2.21 (31 from 14) Peterborough 2 (26 from 13) Southampton 2 (24 from 12) Huddersfield 1.92 (25 from 13) Leyton Orient 1.92 (23 from 12) Rochdale 1.92 (23 from 12) Ultimately, we can only do what we can. 2 points per game is exactly where we should be aiming at over a long period of time (26 games between 1st January and the end of the season), and that's right where we are at the moment. We can only affect the performance of the other teams when we actually play them. In games against the current top 6, we've taken 11 points from 6 games, so not quite up to the 2pts per game target, but a win at Bournemouth next week would fix that in an instant.

There seems to be no consistency to it, one minute it's lightening quick (like loading this thread just then), the next it'll take 10-15 seconds. Haven't noticed it any slower than that though, and no DB errors, so that's a big step in the right direction.

Carlisle

Currently disabled, it's quite a heavy query to run.

The page load speed has been quite variable during the first half (although found loading the match thread much quicker than loading the main board itself), but the speed has really increased at half-time, strangely, when I'd have expected it to grind to a halt with people posting their half-time views. 418 people online at the moment, by the way (329 logged in). Would expect at least 600 around full-time.

Latest update: We are now running the site on the new server, with the database remaining on the old one. So far, everything looks a lot quicker and smoother, but the real test will be when there are 700-odd people online at full-time this afternoon. I haven't made any config changes to the "web server" bit of the new server yet, so it's possible there may be some further adjustments needed, but it's difficult to test that without actually being in the real life situation of loads of people adding posts to the database. I've tested concurrent users online with decent success, but the process of writing to the database is much more resource-intensive than reading from it. In short, we'll see how we go this afternoon...

I think we're back...

Is that an admission that you're intentionally winding people up?

I've heard nothing from the club about it. Last year, they gave those who did it the previous year first refusal, and contacted everyone around Christmas time, so I'm assuming they won't be doing it this time around.

In other breaking news, Titanic has sunk

I've done some further analysis on the database and the actual database is working fine. It's using the query caches quite well and seems to be relatively efficient. However, accessing it with 400-odd concurrent users still has an effect on the original server, so I can now conclude that the problem isn't with the database itself, but with the web server part of the system. As a result, I'm going to switch the codebase and web server across to the new server, and have the DB on the current one. I have no idea how long that will take, and certain aspects of it will require assistance from the hosts, and I don't know if they'll have anyone available for that sort of task this evening.

We probably still got the better end of the deal there

This may take longer than I'd originally hoped

Just as a prior warning, I will be testing various aspects of the new setup this afternoon and evening, so there may be intermittent periods of downtime, particularly of the "Database Error" message variety. Many connection failures will show up instantly (i.e. without the loading page hanging for ages and then deciding it can't connect), this will be where the database process isn't actually running, as when I make changes to the DB configuration, the whole DB needs stopping and restarting. Summary: you will probably see occasional "Database Error" messages between now and midnight. It's not a fault. I'm testing stuff.

If I remember rightly, it was a text-to-enter competition, and the club with the most entries won the £250k. The individual winner was drawn at random from all the winning club's entries, I think.

This is the first time I've come back to this thread since I started it. Bloody hell Memcached is installed, but not activated at this stage. I tried it on our dev site but I couldn't get it working (basically just output a blank page every time!) so put it on the back burner for a while. Once we've got the new DB up and running properly, I'll revisit that as it should provide a benefit for the site. Very good point. I knew about the avatars being loaded via PHP rather than direct from the filesystem, but the thought hadn't crossed my mind that that could probably be made more efficient. It might be an absolute ballache to modify it, mind you, so I might just have a look at image.php itself and add a cache expires header (if there isn't already one - the dateline parameter would need to be removed, mind you)

Evening all As I'm sure you're well aware, we've been experiencing unacceptable amounts of downtime during matches over the last 4-6 weeks. We did think we had solved the problem back in September when a number of system changes were implemented, but it seems as though they only had a temporary effect. As a result, we have decided to change the way the infrastructure is set up. Rather than have the web server (i.e. the bit that handles everyone accessing it and sends the web page back to your browser) and the database on the same server, we are buying a second server whose sole responsibility will be to handle the database. This new server has more than double the power of the current one, and as we believe the main problem has been the resources taken up by the database, it's obviously hoped that lightening the load on the main server in this way will provide a long-term solution. Naturally, while we migrate the database to the new server, the site will experience a further period of downtime, which is obviously regrettable but we'll do our best to limit this to off-peak times as much as possible. We aim to have the new server in place before Saturday's game at Colchester. Baj and I have been as frustrated as you have with the frankly ****-poor performance at crucial times, and hopefully this will correct that. Any (sensible and on-topic) questions, let me know. Cheers Steve Update: 4th March, 3.05pm Just as a prior warning, I will be testing various aspects of the new setup this afternoon and evening, so there may be intermittent periods of downtime, particularly of the "Database Error" message variety. Many connection failures will show up instantly (i.e. without the loading page hanging for ages and then deciding it can't connect), this will be where the database process isn't actually running, as when I make changes to the DB configuration, the whole DB needs stopping and restarting. Summary: you will probably see occasional "Database Error" messages between now and midnight. It's not a fault. I'm testing stuff. Update: 5th March, 11.55am We are now running the site on the new server, with the database remaining on the old one. So far, everything looks a lot quicker and smoother, but the real test will be when there are 700-odd people online at full-time this afternoon. I haven't made any config changes to the "web server" bit of the new server yet, so it's possible there may be some further adjustments needed, but it's difficult to test that without actually being in the real life situation of loads of people adding posts to the database. I've tested concurrent users online with decent success, but the process of writing to the database is much more resource-intensive than reading from it. In short, we'll see how we go this afternoon...

That was probably mine, Matt made a bit of a mess of writing my number on the form

If they've got any common sense (a bit of an assumption, I know...), they'll have been sorting through all the applications as and when they've arrived over the last two weeks or so, working out the average number of away games per application and putting them in order. Then when they start allocating tickets, they just start with the top of the list and work down until there are no tickets left.

My parents bought tickets in the home end with some friends who live round the corner from Dean Court when they first went on sale. For some odd reason, their money was refunded last week. Needless to say, they're not overly impressed.

Definitely, but their fans hate him so there wasn't really ever likely to be a way back for him even after Ferguson was sacked.

I went to Preston v QPR. An interesting game but one lacking in consistent quality from both sides really, as expected the difference was the class of Taarabt and Routledge in the first half. Preston were so poor in the first half that QPR clearly thought they didn't have to bother turning up after the break, which let Preston back into it. Cracking finish for the equaliser by Nicholson, but every other touch he had in the game went straight to the opposition. A very determined second half effort from Preston, but they're so impotent going forward, I can't see how they'll get out of the situation they're in. They brought on an 18-year-old with 8 minutes to go and he had more of an impact than Nathan Ellington, Eddie Johnson and Paul Hayes - their other "strikers" - put together.

Summary, for those of you who can't be arsed to read through the rest of the thread: there is an agreement (which has never been hidden), but the sole beneficiary is this forum. The club receives nothing in return.