On 05/22/2010 04:41 AM, Amos Jeffries wrote:
Edoardo COSTA SANSEVERINO wrote:
Hi all,
I had a look in the archives and the only similar problem I found was
never answered
(http://www.squid-cache.org/mail-archive/squid-users/200801/0152.html) so
I hope someone can help me. I posted this request on
linuxquestions.org but got no reply so I thought I'd be better off
asking you guys ;)
I tried to get reverse proxy working with apache mod_proxy but that
failed so I'm giving squid3 a go but with not much more luck. All
connections to non ssl websites work fine. The following error I
[B]only get the second time[/B] I access the page, the first time the
page is displayed properly! This does not make sense to me but maybe
it will to one of you.
ERROR
The requested URL could not be retrieved
While trying to retrieve the URL: https://deb01.example.com/
The following error was encountered:
Connection to 192.168.122.11 Failed
The system returned:
(71) Protocol error
The remote host or network may be down. Please try the request again.
Your cache administrator is webmaster.
Generated Thu, 20 May 2010 18:58:28 GMT by localhost (squid/3.0.STABLE8)
My setup
--------
+--> (deb02) vhosts running multile http
|
[WWW] -> KVM/SQUID ->+--> (deb01) vhost running a single https
|
+--> (deb03) vhosts running multile http and one
https
My squid.conf
-------------
https_port 443 accel cert=/etc/ssl/deb01.example.com.crt
key=/etc/ssl/deb01.example.com.pem defaultsite=deb01.example.com
vhost protocol=https
http_port 80 accel defaultsite=deb02.example.com vhost
cache_peer 192.168.122.11 parent 443 0 no-query originserver
login=PASS ssl sslversion=3 sslflags=DONT_VERIFY_PEER
front-end-https=on name=srv01
cache_peer 192.168.122.2 parent 80 0 no-query originserver name=srv02
acl https proto https
acl sites_srv01 dstdomain deb01.example.com
acl sites_srv02 dstdomain deb02.example.com second.example.com
http_access allow sites_srv01
http_access allow sites_srv02
cache_peer_access srv01 allow sites_srv01
cache_peer_access srv02 allow sites_srv02
forwarded_for on
---
The first 'successful' connection gives the following entries in the
logs:
-----BEGIN SSL SESSION PARAMETERS-----
MIGIAgEBAgIDAQQCADUEIDrfJnfrcvWw15QVzrwAlKJYsrinM/X+Ge9aeTyO8Fkx
BDBLAPhbkN6LTcdvHMF9YGm8ib5Qwjm05qP3rr7I+LBjpikfjzV5gJSXLfke83U0
ggOhBgIES/WH8aIEAgIBLKQCBACmFQQTZGViMDEucHJlY29nbmV0LmNvbQ==
-----END SSL SESSION PARAMETERS-----
2010/05/20 21:05:21| 192.168.122.11 digest requires version 17487;
have: 5
2010/05/20 21:05:21| temporary disabling (invalid digest cblock)
digest from 192.168.122.11
Normal. Simply means the web server does not understand proxy cache
digest exchange. This can be silenced by adding no-digest option to
the web server cache_peer line.
2010/05/20 21:05:21| fwdNegotiateSSL: Error negotiating SSL
connection on FD 16: error:1408F06B:SSL routines:SSL3_GET_RECORD:bad
decompression (1/-1/0)
2010/05/20 21:05:21| TCP connection to 192.168.122.11/443 failed
[...]
There is your HTTPS problem. Your SSL system libraries are producing
that error when they can't handle the settings.
http://google.com/search?q=SSL3_GET_RECORD%3Abad+decompression
2010/05/20 21:05:21| fwdNegotiateSSL: Error negotiating SSL
connection on FD 16: error:1408F06B:SSL routines:SSL3_GET_RECORD:bad
decompression (1/-1/0)
2010/05/20 21:05:21| TCP connection to 192.168.122.11/443 failed
2010/05/20 21:05:21| fwdNegotiateSSL: Error negotiating SSL
connection on FD 16: error:1408F06B:SSL routines:SSL3_GET_RECORD:bad
decompression (1/-1/0)
2010/05/20 21:05:21| TCP connection to 192.168.122.11/443 failed
<snip storage log>
The second 'failed' connection shows the following log events:
==> /var/log/squid3/cache.log <==
2010/05/20 21:06:11| fwdNegotiateSSL: Error negotiating SSL
connection on FD 15: error:1408F06B:SSL routines:SSL3_GET_RECORD:bad
decompression (1/-1/0)
[...]
2010/05/20 21:06:12| fwdNegotiateSSL: Error negotiating SSL
connection on FD 15: error:1408F06B:SSL routines:SSL3_GET_RECORD:bad
decompression (1/-1/0)
2010/05/20 21:06:12| TCP connection to 192.168.122.11/443 failed
2010/05/20 21:06:12| fwdNegotiateSSL: Error negotiating SSL
connection on FD 15: error:1408F06B:SSL routines:SSL3_GET_RECORD:bad
decompression (1/-1/0)
2010/05/20 21:06:12| TCP connection to 192.168.122.11/443 failed
<snip storage log>
store.log is irrelevant to most uses. You can safely set it to "none"
in your squid.conf file.
Any help would be greatly apreciated.
As a side note. If anyone can tell me how to show the IP of the
squid server rather than the internal IP of the webserver (as in the
error) that would be a bonus ;)
The error is correct.
The link client->squid is not working perfectly.
The link squid->server (via internal IPs) is failing.
Thus you get a report telling you which link out of the two has
failed. Changing that will only make you look in the wrong place for some
Amos
Hi Amos,
Thanks for your answer but I am a bit confused now.
Are you saying it's an SSL problem?
There is your HTTPS problem. Your SSL system libraries are producing
that error when they can't handle the settings.
http://google.com/search?q=SSL3_GET_RECORD%3Abad+decompression
Or are you saying that squid is unable to forward SSL to an internal IP?
The link client->squid is not working perfectly.
The link squid->server (via internal IPs) is failing.
Is squid the tool for me if I want to implement the above setup? What
would you recomend, Squid, mod_proxy, ...?
Thank you for any help or pointers to documentation you might come up with.
-Edoardo
