|
> I meant the content of the service being offered, eg the html files
> on the realservers.
Ahh ok, the site is mirrored across the two servers using rdist. Session
management and content is database driven, so a client can transparently
move from one server to another without having any problems. I really don't
need persistence, but it's there just in case until I'm able to assess
everything a bit better. I did a few months of fiddling around in testing
without any problems, but apparently I forgot to check the upload stuff :-/
> > HTTP, like it says in the subject ;).
>
> OK I have heard of it then :-)
>
> I'm not familiar will uploading with http though.
It's just a type of HTTP POST.
> > Its not just image uploads that cause the server to hang, large HTTP
GETs
> > are also problematic.
>
> I expect it will be easier to test if this is fixed and you should try
this
> first.
After a bit of testing, a 934 byte file to POST (upload) works, but a 935
byte file consistently DOES NOT. What does this mean? I've never seen a
limit like that. Am I leaning up against some weird packet fragmentation
problem (even though the tcpdump stuff shows DF)?
> I don't see anything odd in your tcpdumps.
Hooray! ;-)
I've got to get this little problem fixed by Tuesday am, one way or another.
I have a hunch that this little issue is causing problems with logging in
for some people since that is done with a HTTP POST over https. Next step
is to test if this issue exists on the other realserver or not. Maybe a
route or interface conf issue?
Thanks,
-Jacob
|
