h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

I wounder...

Well, things have taken a turn for the worst at the moment! Now Greebo is experiencing the same problem with her personal and Post account - that makes all the editorial team of <./>ThePost</.> unable to update or add entries.

shazz

This isn't a virus is it? I can't think of any other reason why Greebo should suddenly 'catch' it?

*sprays disinfectant around*

This is a worrying development .

I'm still around though. Mind you, I haven't tried long posts for a while...

I'm afraid I don't have any ideas. I take it Jim Lynn couldn't come up with anything either?

Reading Greebs' thread on this subject it seems that her problem is not *exactly* the same though. Am I right?

Well, that makes three then because I can't send long posts either.
Jim has already told me that it is an h2g2-problem. He wants me to run a test over here but unfortunately I have been unable to find the time. I will run it tomorrownight and send him the results.

Greetz,

TM.

But Greebs' problem isn't with the length of posts - she's not getting the edit dialog at all. Or have I misunderstood?

Amy

sounds like trouble.

-- DoctorMO --

Now that was weird. I put in a long entry under the Post, and it all seemed to go fine, but then, when I deleted it, I was suddenly logged out?!?! and had to log back in again.

I had misunderstood. Greebs' problem is with long posts and long Guide entries (more than 23 lines). Perhaps she has caught something from Shazz . Perhaps we should consult a vet .

If she's caught something from shazz, it may well be too late for a vet

"Jim has already told me that it is an h2g2-problem. "

What I actually said was that it looked like it was a problem at our end. However, when I looked more closely, it seems that the problem still originates at your end. We *are* receiving the requests (the long ones which fail) and our logs are showing a very long timeout, but what I *think* is happening is that the request is getting mangled somehow on the way to us, because our logs are telling me that we're only receiving 87 bytes in the request (as opposed to 20779 in a request that worked) which is definitely not right. The timeouts are possibly because we're waiting for the rest of the request to come in, and giving up after about 400 seconds when it doesn't.

So, sadly, the problem isn't specifically at our end. I can't rule out a proxy server mangling the requests, but we're definitely not seeing the whole request at our end.

I'll ask around to see if it could be a proxy related problem, but for now it still looks to me like it's coming from your end.

Correction - I said 'it's definitely a problem at our end', not 'probably'. At the time I had spotted the log entries that had the long timeout, and had assumed that there was something in those requests causing DNA to process forever. After I sent that email, and had looked in vain for these requests in our secondary logs (logs written by DNA directly rather than the raw web server logs) I then noticed the short length of the request, which now suggests something has mangled the request before I see it.

I just wanted to clear that up, in case TM thought I was denying what I said - I just hadn't re-read the mail I sent until now.

Didn't d'Elaphant have a similar problem with his router which he solved, I think, by downloading a patch?

Well Jim, all I can say is that I'm glad other people are apparently also experiencing this problem now. Sorry if that sounds mean but - as I am not exactly an network-specialist - I'd rather see the problem on the 'BBC-side' than on mine. Anyway, I will send the data you requested shortly and then start praying tot BOB for a solution. I know one thing for certain though : if this problem persists there must be something special about the h2g2-software(not that I doubted that for a second of course). All the other forums work fine for me.

Oh well, I'm just going to keep on hoping.

Greetz,

TM.

Actually, it can still be a problem at your end. If the problem is (as someone else suggested) a problem with the setting of the MTU size in your router or OS, and the router which is choking on the packets is on the route to the BBC but not those other forums, then you'd get exactly the result you are experiencing.

The fact remains, our servers (i.e. the h2g2 servers) are *not* seeing complete requests from you when it fails, but we are when it succeeds, which tells me it's happening further up the chain, but how far is hard to see.

It's also very odd that the three people reporting the problem are all working on the Post. Are you in the habit of running software to tweak your dial-up networking performance? I've come across references to programs which tweak DUN settings which can cause a problem very like this.

Nope, we don't use any special tweaking-software. Just plain vanilla Windows2000 in our case(Shazz and me). If you give me a new value for the MTU I can try that tonight. Would 1440 do ?

TM.

I managed to add all ten of the next editions of The Post with no problem at all on Friday - but then they are very short additions. When I tried to add the latest Back Issue page I could only manage half of it and had to send the rest of the coding to Greebo to put in later.

This, as you know, isn't just happening with The Post Team but with our personal accounts as well - it's just a nasty coincidence I'm sure.

I too have had that problem once Pastey - being automatically logged off when I deleted a test page.

shazz