The problem is that I cannot download a file from another server under these circumstances:
* My site running ssl
* Proxied site does not seem to matter (tested both http and https)
* Using Internet Explorer 6 (Mozilla works!)
* Rewrite URL:s set to "yes"
* The targeted file is a binary file - I've tested .txt(ok) .html(ok) .doc(failed) .pdf(failed)
At the non-webgui site there are a link to a pdf doc. When I open this I get a file download prompt and:
"Internet Explorer cannot download aaa.pdf from our.server.com
The file could not be written to cache."
This is an IE only error and I googled and found following:
"In SSL, Explorer does not like the Pragma: no-cache header. That was the problem"
So the Pragma: no-cache is set in Session/Style.pm
Snapcount :"I think the discrepancy happens because proxy uses LWP to request the content and I'm guessing by default it sends that cache pragma in the header. We could probably tell LWP not to do that"
I refer to support case: http://www.plainblack.com/plain_black_support/tricky-ssl-problem/
/Erik
ALSO, When you fix this bug, please do something about the way http proxy handles large files.
Have it first find out how big the file is, then if its small do not change anything. BUT , if its a large file then you should store that in a temp file rather than in memory because its causing webgui and the entire server to seize and lock up. THANKS,
grnoc
"Do something about the way HttpProxy handles large files..." is an RFE, and one that would be difficult to implement. Far better to set up a different apache instance.
The HttpProxy is not meant to allow an entire website to be used inside of a WebGUI style, at least not without some minor modifications.
As an aside, please keep threads on-topic. Even if it was a bug, your post is completely different from the first post (even though they share a symptom).
I could not reproduce this on IE7 using the following link:
http://demo.plainblack.com/demo1167851842_767/home/untitled
Please visit that link and tell me what you see.
That is proxying this file: http://demo.plainblack.com/demo1167851842_767/uploads/SW/FV/SWFV0DOuZGYqi61THU1QKQ/Chouchin-Reinensai-Lantern.gif
This is the case:* Internet Explorer 6
1. HTTP -> HTTP - OK2. HTTP -> HTTPS - OK3. HTTPS -> HTTPS - FAILS
I cannot reproduce this on demo site since https is not active.
Iget this kind of error as well:
The proxy hangs (httpd process) when I download files using HttpProxy (FF and IE 7). This is not big files (max 1 mb) but still causes our server to peak at 99% cpu.
From what I've been reading out on the internet tubes, you're right that this bug seems to occur because of the "pragma: no-cache" HTTP header.
Since I can't figure out a good way to reproduce this, is there any way for you to give me the full HTTP response for the HTTPProxy request for the file?
Usually using `wget --save-headers <link that causing problems>` will work.
It might be possible to workaround that issue with IE.
Another thought: Can you reproduce on IE7?
