Something like a year ago, I set up mod_rewrite to automatically redirect all http requests to https. That's been working fine and still is. There are a few cases where I still need to permit http access, so I have a few extra rules to override the main one for these special cases. I added a new one recently and found it isn't working. The rule is working and the http requests are getting through without being redirected, but in many cases only headers are being sent back with no data. curl -v -f is showing no data and then failing with error 18 (which I believe means the amount of data sent doesn't match what was reported in the header - and it seems to be no data sent at all causing this). Whether this failure occurs seems to depend on where I make the request from and how big the file is. Requests from the server itself (but via the external interface) mostly work. From other machines on the network, the bigger the file the more likely is failure. This suggests timing may be involved to me, but I have no other evidence that that is the case.
I see no errors in the error logs. All the accesses show up in the access logs (failure or not). I've tried changing places in the config where log level is defined to "debug", but still nothing is reported.
I tried removing my mod_rewrite rules completely and then I see problems throughout the sites I serve when using http rather than https. I have horde/imp webmail: strangely the php files are still interpreted and the result served, but css files and images don't appear.
I tried removing mod_ssl and mod_reqtimeout, but that seemed to make no difference.
I just wondered if anyone could suggest a possible cause, or a way to discover more information on the problem. Any help would be much appreciated.
Cheers, Paul. --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
|