[SOLVED] OTRS Appliance 4.0.6 - "Impossible to send message to"

[andrewn]

Alright, I'm hoping you guys might be able to help me out on this one. I see that it's a common issue, but so far I haven't been able to find a solution.

I downloaded the OTRS Appliance (4.0.6) earlier today and installed the ITSM 4.0.6 module as well. Everything was going fine until I went to try and create a new ticket. I tried to send it to my email (andrew@t*************.com), but it gave me an "Impossible to send message" error. I googled for a few hours, found quite a few other threads on the issue, but it seems most of what I found was outdated or related to bugs that were previously patched.

I have tested with Sendmail, SMTP, SMTPS, and SMTPTLS and I always get the same error:

Code: Select all

Oops! An Error occurred.
Error Message: Impossible to send message to: andrew@t*************.com .
.
You can  Send a bugreport or go back to the previous page.

Code: Select all

Error Details
Backend ERROR: OTRS-CGI-41 Perl: 5.10.1 OS: linux Time: Fri Mar 13 18:35:47 2015 Message: Impossible to send message to: andrew@t*************.com . RemoteAddress: XX.XX.XX.XX RequestURI: /otrs/index.pl Traceback (1838): Module: Kernel::System::Ticket::Article::ArticleSend Line: 2560 Module: Kernel::Modules::AgentTicketEmail::Run Line: 1427 Module: Kernel::System::Web::InterfaceAgent::Run Line: 987 Module: ModPerl::ROOT::ModPerl::Registry::opt_otrs_bin_cgi_2dbin_index_2epl::handler Line: 41 Module: (eval) (v1.99) Line: 204 Module: ModPerl::RegistryCooker::run (v1.99) Line: 204 Module: ModPerl::RegistryCooker::default_handler (v1.99) Line: 170 Module: ModPerl::Registry::handler (v1.99) Line: 31

The only thing I get in /var/log/messages is:

Code: Select all

Mar 13 18:35:47 localhost OTRS-CGI-41[1838]: [Info][Kernel::System::Email::Send] Error sending message

The error message is consistent no matter what mail configuration is set.

I have verified my SMTP credentials using outlook and thunderbird, and everything is good to go. I can connect and send through either of these applications no problem.

My settings are currently:

• SendmailModule: SMTPS
• SendmailModule::CMD: /usr/sbin/sendmail -i -f [DEFAULT]
• SendmailModule::Host: mail.t*************.net
• SendmailModule::Port: 465
• SendmailModule::AuthUser: alerts@t*************.net
• SendmailModule::AuthPassword: ThisIsDefinitleyNotARealPassword1337
• SendmailBcc: <blank> [DEFAULT]

All other settings under Core::Sendmail are blank and unchecked by default.

This is extremely frustrating... One would expect an appliance to be good to go straight out of the box, but I suppose not. The appliance is in a VMWare VM, and has plenty of resources. This is actually the same VMWare server that hosts my mail client, so I know there are no network-level issues.

Any help you guys can provide would be greatly appreciated. I would really like to move to OTRS fairly soon (and get away from OSTicket), so the sooner I can get this resolved the better. Otherwise, I haven't seen any other issues crop up (yet).

Thanks!

[crythias]

Impossible to send message means smtp service actively refuses. This is not related to OTRS being unable to do this out of the box.

[andrewn]

Why would it be refusing though? All my settings are valid (tested multiple times) and I've even tested with different providers (e.g. Gmail). I've even tested using sendmail, but that didn't work. Any ideas?

[andrewn]

SOLUTION
I stumbled across this Wiki page which suggested using the Admin Notification page in the Admin panel to test mail sending, rather than creating numerous new tickets like I was doing. I decided to test this way for grins after verifying settings again, and I got a new error stating that it was "impossible to connect" to my mail server. Naturally, I logged in and tried to ping my server, and the name didn't resolve! After a few minutes of head scratching, I went over my network settings again, and found the issue in my DNS configuration.

It turns out that the appliance by default uses what appears to be Pascom's internal DNS servers. While logged in via SSH as root, I edited /etc/resolv.conf, removing the original contents and replacing it with:

Code: Select all

nameserver 8.8.8.8
nameserver 8.8.4.4

Once resolv.conf was updated, I was able to send a mail no problem.

Could have sworn I set this during the install and/or via the system config page screen at the terminal, but it's like it didn't take. I'm going to run through the install again on a test box just for grins sometime and see if this was something I missed, or if there is a bug. Not sure when, but one of these days.