On 18/02/2012 8:20 a.m., ALAA MURAD wrote:
> Dear Amos ,
>
> Thanks so much for your help, I hope this changes makes more sense. I
> have been following your comments and I feel squid is running better
> than before.
>
>
> The setting now feels more cleaner and squid out less errors, but
> mainly now I'm suffering from one thing, sometimes , I get this error
> :
> clientNegotiateSSL: Error negotiating SSL connection on FD 355:
> WSAENOTCONN, Socket is not connected. (10057)
> clientNegotiateSSL: Error negotiating SSL connection on FD 356:
> Resource temporarily unavailable (11)
The TCP connection "FD 355" is closing before SSL details can be
negotiated between Squid and the client.
The TCP connection "FD 356" is doing negotiation but was unsuccessful
due to "resource unavailable" SSL error. Probably broken TCP again, or
SSL security system failing to access things.
Are you able to identify whether those are connections between the
peers? or from outside clients?
>
> Both servers are connected peer-to-peer (back-to-back) and a running
> ping confirms connection is up all the time!
>
> One thing to note about the above problem, that this mainly causing
> errors in IE , saying that the certificate has expired (this is really
> random and not sure what is wrong with it, waiting few minutes
> certificate is OK again ! ).
>
>
> Other than that, it's perfect !
>
>
> Also I did those :
>
> 1- I have removed all http (port : 80 ) as this reverse proxy running
> only on SSL. Also do I need all this Safe_Ports for a site that only
> serves port 433 ?!
Safe_ports is optional for a reverse-proxy. It is only relevant to
forward-proxy and management port traffic.
if you chose to do the management port, IF you choose to use one (the
management requests can be done over accel ports too).
> 2- Removed cache& unwanted logs .
> 3- Still confused with defaultsite=www.eservices.mycompany.com , I
> kinda got the point , but not sure what is the perfect alternative
> (removing it caused a header error in browser ).
The vhost option tells Squid to use the client Host: header to identify
the domain. defaultsite= is a backup for use if the client did not send
Host: information at all.
> 4- Also my rule "http_access allow all" I guess that is needed in
> reverse proxy as I want to allow all to hit my site.
No. Allowing your site is all that is needed.
You earlier had "http_access allow mycompanyserver" which was doing the
correct thing, and doing it before the Safe-ports and surrounding access
controls. That (at the top) was the right way to do reverse-proxy access
controls, so the forward-proxy ones do not get a chance to slow down the
requests.
"allow all" has the effect of passing all requests on to the backend
server. By only allowing your site requests to go back to the server
Squid can protect against DDoS with randomised domain names pointing at you.
> 5- The redirector, mostly output blanks (99%) , and in rare event it
> intercept and rewrite the url.
Okay. good.
> for concurrent rewrite isn't why we can
> load many helpers and that will help in concurrency? I'm good in
> Threading in Java, but what I'm afraid of is to confuse squid , when
> printing stuff out of order in multi-threaded application.
The concurrency works by Squid sending each request with a channel-ID
token. The redirector can send its reply back at any time with the same
ID token and Squid does not get confused. This saves memory and CPU
running many helper processes, when a few multi-threaded ones can be
used. Even without multi-threading it raises the amount of requests each
helper can service, saving memory and user annoyance.
> 6-Not sure about this as it's a windows server , "refresh_pattern -i
> (/cgi-bin/|\?) 0 0% 0"
Windows has nothing to do with refresh_pattern. This is a pattern in the
URL indicating a dynamic script is handling it. You can add
\.asp|\.php|\.chm etc there too if you know a script is not generating
correct cache-controls and also not using '?' in the URI.
> 7-finally , I wasn't able to get squid to work without
> 'sslflags=DONT_VERIFY_PEER"' , what is the impact ?
Without it Squid attempts to validate the peer SSL certificate against
the root CA Squid (via the openssl library) trusts. Squid is not able to
perform a certificate popup like browsers do. So is forced to abort the
connection setup whenever validation fails.
If you have self-signed the backend server certificate you can use that
flag, or you may be able to import your self-created root CA into the
set trusted by Squid. See openssl library documentation for how to do
that, it may or may not be possible and I have never tried it myself.
>
> Once again, thanks for your help. You have no idea how much this is
> helping me :)
You are welcome.
>
> Best Regards,
> Alaa Murad
>
> ----------------------------------------------------------------------------------------
>
> https_port 443 cert=C:/Interceptor/cert/baj.cert
> key=C:/Interceptor/cert/baj.key
> defaultsite=www.eservices.mycompany.com
> cache_peer 192.168.1.2 parent 443 0 no-query no-digest originserver
> ssl sslflags=DONT_VERIFY_PEER name=bajsite
> redirect_children 25
> redirect_rewrites_host_header on
> redirect_program C:/java/bin/java.exe
> -Djava.util.logging.config.file=C:/Interceptor/redirector/RedirectorLogging.properties
> -jar C:/Interceptor/redirector/Redirector.jar
> logformat common %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %Hs %<st %Ss:%Sh
> logformat combined %>a %ui %un [%tl] "%rm %ru HTTP/%rv" %Hs %<st
> "%{Referer}>h" "%{User-Agent}>h" %Ss:%Sh
> access_log c:\squid\var\logs\access.log common
> cache_log c:\squid\var\logs\cache.log
> cache_store_log none
> refresh_pattern ^ftp: 1440 20% 10080
> refresh_pattern ^gopher: 1440 0% 1440
> refresh_pattern . 0 20% 4320
> acl all src 0.0.0.0/0.0.0.0
> acl localhost src 127.0.0.1
> acl to_localhost dst 127.0.0.0/32
> acl SSL_ports port 443 # https
> acl SSL_ports port 563 # snews
> acl SSL_ports port 873 # rsync
> acl Safe_ports port 80 # http
> acl Safe_ports port 81 # http
> acl Safe_ports port 21 # ftp
> acl Safe_ports port 443 # https
> acl Safe_ports port 70 # gopher
> acl Safe_ports port 210 # wais
> acl Safe_ports port 1025-65535 # unregistered ports
> acl Safe_ports port 280 # http-mgmt
> acl Safe_ports port 488 # gss-http
> acl Safe_ports port 591 # filemaker
> acl Safe_ports port 777 # multiling http
> acl Safe_ports port 631 # cups
> acl Safe_ports port 873 # rsync
> acl Safe_ports port 901 # SWAT
> acl Safe_ports port 8080
> acl CONNECT method CONNECT
> http_access allow to_localhost
to_localhost is designed to catch attacking requests (requests designed
to loop inside Squid until all TCP ports are consumed). It shoudl be
used in a deny line.
Yours is also missing the "0.0.0.0/32" IP pattern.
> http_access deny !Safe_ports
> http_access deny CONNECT !SSL_ports
> http_access allow localhost
> http_access allow all
> http_reply_access allow all
> coredump_dir /var/spool/squid
> forwarded_for on
> redirect_rewrites_host_header on
> buffered_logs on
> never_direct allow all
> cache deny all
Amos
Received on Sat Feb 18 2012 - 06:40:59 MST
This archive was generated by hypermail 2.2.0 : Sat Feb 18 2012 - 12:00:03 MST