Problem sending emails only from Outlook, everything else works fine

[gbarchi]

Hello,

I am having a bit of a problem sending emails from Outlook. This was working before a reboot of the server we had to do this morning.

Right now webmail works, both sending and receiving, Blackberrys work using IMAP 993. POP without SSL, using 110 works using Outlook. But sending email via SMTP using 587 without TLS is not working.

Release 7.1.3_GA_3346.UBUNTU10_64 UBUNTU10_64 FOSS edition.

The log shows

Nov 14 11:12:22 mail saslauthd[14638]: authentication against url 'https://mail.domain.com:7071/service/admin/soap/' caused error 'curl_easy_perform: error(56): Proxy CONNECT aborted due to timeout'
Nov 14 11:12:22 mail saslauthd[14638]: url 'https://mail.domain.com:7071/service/admin/soap/' will not be used for (at least) 600 seconds
Nov 14 11:12:22 mail saslauthd[14638]: Authentication cycle re-elected url https://mail.domain.com:7071/service/admin/soap/, giving up ...
Nov 14 11:12:22 mail saslauthd[14638]: auth_zimbra: gmartillo auth failed: curl_easy_perform: error(56): Proxy CONNECT aborted due to timeout
Nov 14 11:12:22 mail saslauthd[14638]: do_auth : auth failure: [user=gmartillo] [service=smtp] [realm=] [mech=zimbra] [reason=Unknown]
Nov 14 11:12:22 mail postfix/smtpd[8315]: warning: unknown[195.100.100.45]: SASL LOGIN authentication failed: authentication failure

I see that there is a problem connecting to the service, but not sure why because from the same box I can wget to the address or any other address without problems or timeouts.

What could have changed after the reboot?

Hope someone can help me.

GP

[gbarchi]

In case this helps, I do notice it is connecting...

Zimbra is on 56, I am using 57 for testing.

tcp6 0 0 195.100.100.56:7071 195.100.100.57:10809 ESTABLISHED

tcp6 0 0 195.100.100.56:7071 195.100.100.57:10810 ESTABLISHED

[gbarchi]

In case someone asks....

"Domain" stands for my current domain.

dig mail.domain.com


; <<>> DiG 9.7.0-P1 <<>> mail.grasasunicol.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 22022
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 0

;; QUESTION SECTION:
;mail.grasasunicol.com. IN A

;; ANSWER SECTION:
mail.grasasunicol.com. 86400 IN A 195.100.100.56

;; AUTHORITY SECTION:
grasasunicol.com. 86400 IN NS 195.100.100.56.

;; Query time: 11 msec
;; SERVER: 127.0.0.1#53(127.0.0.1)
;; WHEN: Mon Nov 14 18:52:47 2011
;; MSG SIZE rcvd: 83

host $(hostname)
mail.domain.com has address 195.100.100.56

[gbarchi]

SOLVED.

The problem was we have just implemented Endian as a UTM solution (about a month ago). The restart of the Zimbra server kicked in a https_proxy configuration that had not come into effect before, and did today after the server was rebooted.

The bottomline was that Zimbra was trying to make connections over port 7071 which is not regularly allowed to go through squid. The solution was to eliminate the https_proxy and Outlook was able to send emails once again. This is the best solution because it doesn´t make sense to send this traffic through squid when it is all local traffic.

Hope this helps someone in the future.

GPB