Thread

For several months now, I have been noticing some odd hiccups when reading the expressionengine.com forums. The problems seem to be peculiar to this site. I am using Firefox 3.0.x and 3.5 on both Windows and Linux, and I see these things on all of them.

1. Often (say 30% of the time) when going to http://ellislab.com/forums for the first time in a browsing session, Firefox will just say it’s waiting for the site, and wait and wait, and the opening page never shows up, even after 30 seconds or so. If I reload the page, it loads promptly.

2. From time to time (every 50-100 page accesses), the CSS seems to be messed up. The usual symptom is larger than normal text, and sometimes odd colors. Again, a reload fixes it.

I should enable FireBug and see if something’s not fetching properly, but so far I haven’t remembered to do this beforehand.

Anyone else seen this? At first I thought it was some minor Firefox issue, but I almost never have the same symptoms on other sites, so I wonder if it’s a hosting issue instead.

(Not sure where to post, since it’s a site issue, not an EE issue.)

Dan

Never seen this, and if there were glitches like this, I’d see ‘em. I’m on Safari though - what you describe makes me think of caching behavior, not necessary browser but possibly proxy caching.  What ISP are you on, and are you behind any sort of company networks?

It feels to me like some kind of incomplete fetch problem: like an HTTP request never finishes or gets truncated.

My home ISP is Verizon (FIOS), and I am not using any proxy. But I’ve also seen this at two workplaces, one with no proxy and one with a very vigorous proxy.

I just tried to duplicate this with Firebug on in a fresh browser session with a cleared cache (and cookies), but could not. But I’ll try to remember to turn on Firebug before browsing the forums in the future and see what I can find out.

I definitely experience problem #1 at least a few times a week.  From FF or Safari (mac).  It has happened to me from the office (DSL), home (cable), as well as my realtor’s office (I was bored while she was typing a contract).

Attached is an example of the font size oddity (normal version is on right). It is not always just the numbers. Firebug’s idea of the CSS at this point did not seem damaged.

I see these problem 1 on occasion as well. Sometimes, the page will load without the CSS. As you said, a refresh clears it up. I’m using Firefox. I always assumed it was a temporary server glitch.

I see #1 from time to time, too (Safari and Firefox, Mac). It’s as if the request just ‘hangs’ and I get a white browser screen. Refresh takes care of it. Never had #2 happen, though.

Add another squirrel to the EE Hosting engine and all will be better.

😉

i see problem #1 all the time, at least daily on EE site. Typically when i load the EE home page as this is usually my entrance page. Typically this does not happen for me until the late afternoon UK time, so when US comes on line!
I have always assumed it was a dropped request at busy times - somehting i have occasionally seen on other hosting services. But as i look at the responces now i see that many appear to be from the UK, and wonder if it is more of a network issue across the pond!

But as i look at the responces now i see that many appear to be from the UK, and wonder if it is more of a network issue across the pond!

Good call. I’m sitting on the other side of the other pond. Where are EE’s servers? Minnesota?

4,022 miles from MN to London. 3,965 from MN to HNL.

Two words: Slow packets.

But I, the original poster, am in Massachusetts, with good connectivity from all the places I’ve seen the problem. So for me it that is not the issue.

Well I can say for sure its not server or network resource limits or issues at play on this end,  so has nothing to do with resource availability.  The odd thing is that we run remote URL testing on all of our company home pages from testing locations in the US, Canada, Europe and Asia every 1 minute.  Not only is this to test our site availability but also gives a picture of global location http response times.  While these tests would likely not “see” a css issue, they would for sure pick up any failed page load requests as they require full html, and look for known response strings in the footer of the pages being tested.  This is outside the overall health monitoring of our clusters of servers, and network infrastructure as a whole too.

Will be doing more of an investigation into this, though this could easily be a specific network path outside of our direct control, ie: certain router/provider, possibly passing through Chicago (major midwest routing point a lot of providers run routers through), etc.  Seems equally as weird that its only ever your first request but an immediate second request would be normal, as well as no reports from the always online staff like Lisa 😉

Thank you for the reports all, and we will see if we can track this back to any upstream concerns.

I get #1 quite often too, though I think it is something to do with cookies.

(I’ll leave aside the CSS issue, which may be something else completely.)

It seems possible my problem #1 is some lifetime issue, either caching or cookies, as I cannot recall having this problem except during the first access of the day. Daniel Walton’s post might be a hint.

Clearing the cache and cookies does not reproduce the problem, so perhaps it is an interaction between an expired and a non-expired thing, or that something must be present but expired.

I see “mintyfresh.js” and some other interesting lifetime-related stuff going by.

If Nevin’s URL testing does not save cookies or do any caching and/or doesn’t run Javascript, then it may not catch this. Nevin, from your description it sounds as if it you fetch all the components of a page, is that right? (E.g., the EE home page makes several dozen requests.)

Perhaps another question is whether anyone sees this on their own EE sites ever? I assumed it was not an EE issue, but perhaps it’s a weird interaction.

Anyone seeing this on Safari, open Window > Activity and see if any resources failed to load, and if so make note of them.  Also when it occurs open Develop > Show Error Console and see if any errors are logged.

i would like to add though… this is not somehting i have been experiencing of new… it has been happening for a long time for me, but never worried me as a refresh saw your site load perfectly. I always suspected it being a UK US issue, or somehting in my more local area such as DNS etc.