Well I’m a bit stumped! Since weather-watch.com was moved to the new server, I’m having issues with the speed of the forums when viewing them under my trusty old Mozilla 1.7.8/Fedora Core 2 combination.
I’ve never struck this issue before, until the forums were moved to the new server (co-incidence?) and I started to notice the Forums taking a long time (10-20 seconds) to display after clicking on one of the Forums. The page just sits for ages with nothing but the Titlebar visible (see screenshot). Eventually the forum of choice appears, but it’s incredibly frustrating to use.
I logged out, and the problem immediately went away and I thought, aahhaa, clear the cookies and cache. I did that, and restarted Mozilla. I know the cookies were cleared as I had to log back into all the other sites I have cookies set for, including weather-watch.com.
What is really weird is that the problem is not apparent under Opera, only Mozilla (which is my “home” browser).
Does anybody have any thoughts as to why this might be happening. I definitely cleared both cookies and cache (multiple tries) and this did not solve the issue. I’ve tried logging out/in numerous time but nothing seems to affect the problem. As soon as I log in, I’m back to snails speed getting the forums to display in Mozilla.
Strange it’s only after you login to forum. Some kind of authentication or trusted site overhead? Maybe try a traceroute when logged in and not logged in and see if there is any difference? FWIW, I have not noticed any problems here, might even be slightly faster since the move.
For me it’s been considerably faster since the move.
Thanks for the thoughts guys, I’m tempted to think it’s cookie related, due to the fact it’s only when I’m logged in. I don’t think it’s a latency or server performance issue, I think it’s something a little weirder than that, like how Mozilla is handling the cookie. It’s just weird that it’s only started to become and issue since the server move.
I wonder if Chris has made any changes to the SMF software and/or authentication mechanism. Perhaps it was something related to the Bug Tracker login? Either way, my Mozilla setup isn’t handling it very well any more.
The Bug Tracker login fix is in the Bug Tracker code and won’t affect the forum operation in any way.
The only thing I’ve done recently is turn on HTTP compression, but that should be negotiated between the web server and browser and if your browser can’t handle it then it shouldn’t be enabled for your session. However, compression is done by the web server software and that doesn’t have any concept of logged in/out of the forum, so the problem would appear all the time if this was the problem.
The new server will have a newer version of Apache, but I can’t see that being the problem. I’m not surewhat else could be different between the logged in/out states.
Thanks for your thoughts and info on that Chris. It has got me a bit stumped. I think the only thing left for me to do is run some ethereal packet captures tomorrow and see if I can identify anything funny going on in the TCP stream (particularly any differences between the logged in and guest states). It’s the only thing I can think to do. It has to be something unique to Mozilla (or at the very least this version I’m running) as Opera doesn’t have a spot of bother and obviously no body else has reported this which points to a problem at my end (I certainly haven’t made any changes to the config of my system for months but something has upset the space time continuum!) Maybe my PC is trying to tell me not to spend so much time on the Weather-Watch forums!?
I’ll let you know if the packet captures turn up anything.
JaxWeather has found this which might be significant considering the version of Mozilla you’re running… http://weblog.philringnalda.com/2004/06/08/gzip-well-sometimes-its-simple although I can’t work out why this would be different when you’re logged in.
I wouldn’t spend much time looking at packet captures (unless that kind of thing turns you on ) until I’ve had a chance to look at compresion/character sets.
Thanks for your efforts on this Chris! I’m please to report I found a solution/workaround/potential cause. I turned off Pipelining under the Advanced HTTP Networking settings of Mozilla and taddaaaa, problem disappeared. Not sure if that gives you any extra clues, but at least I’ve found a workaround in the meantime (not sure how much it will impact the rest of my browsing performance having turned it off but time will tell).