Issue Mail AutoConfig not working, returns 400 Bad Request

[TimReeves]

This new feature is attractive, so I wanted to activate it. I did, but Thunderbird was obviously not getting the data.
I debugged the feature, trying to call the URL which Thunderbird would use: 400 Bad Request.

I tried this on 2 different servers, one with an external FW in front of it, the other just a normal vServer at a major hoster. Same in both cases.

Both servers run debian 9 and have Plesk Web Admin Edition - could that be a problem?

I do use custom/domain templates (mainly to get rid of all that unused Apache Config), so I tried reconfiguring using the default templates, and temp. removing my extra Nginx config - still the same problem. I don't know how to move forward.

Here is the relevant nginx config as generated:

# mailconfig
location ~* ^/autodiscover/autodiscover\.xml$ {
try_files $uri @mad;
}
location ~* ^(/\.well-known/autoconfig)?/mail/config\-v1\.1\.xml$ {
try_files $uri @mad;
}
location @mad {
rewrite ^(.*)$ /mailconfig/ break;

proxy_pass http://127.0.0.1:8880;
proxy_set_header X-Host $host;
proxy_set_header X-Request-URI $request_uri;
}
# mailconfig

My "Apache & nginx Settings" have ALL the checkboxes for Nginx OFF.

Help would be much appreciated!

 

[TimReeves]

Well now I'm bamboozled: I tried adding a new account in Thunderbird anyway, and it looked like TB had got the data. So I checked the TB console log, and yes:

2019-11-07 18:21:57 mail.setup INFO call 1 took 130ms and succeeded at <https://(redacted).space/autodiscover/autodiscover.xml>
2019-11-07 18:21:57 mail.setup INFO call 1 took 141ms and succeeded at <http://(redacted).space/.well-known/autoconfig/mail/config-v1.1.xml>
2019-11-07 18:21:57 mail.setup INFO status msg: Einstellungen wurden bei Ihrem Anbieter des E-Mail-Diensts gefunden

The exactly same URLs (copy-paste) entered in a Browser (FF and Chromium) fail with "400 Bad Request".

Obviously there is something happening here that I'm missing - can someone please enlighten me?

Thanks,
Tim

 

[VicenteS]

I have the problem too with Plesk Obsidian 18.0.24

Every check returns Error 400 (not 404), I have tried:

- Direct access
- Thunderbird
- Outlook
- Microsoft Remote Connectivity Analyzer

I have activated debug log and is this:

[Wed Mar 11 11:14:31.808548 2020] [ssl:debug] [pid 6902] ssl_engine_kernel.c(366): [client *.*.*.*:53914] AH02034: Initial (No.1) HTTPS request received for child 0 (server ***.com:443)
[Wed Mar 11 11:14:31.808685 2020] [authz_core:debug] [pid 6902] mod_authz_core.c(835): [client *.*.*.*:53914] AH01628: authorization result: granted (no directives)
[Wed Mar 11 11:14:31.808768 2020] [proxy:debug] [pid 6902] mod_proxy.c(1229): [client *.*.*.*:53914] AH01143: Running scheme http handler (attempt 0)
[Wed Mar 11 11:14:31.808788 2020] [proxy_fcgi:debug] [pid 6902] mod_proxy_fcgi.c(913): [client *.*.*.*:53914] AH01076: url: http://127.0.0.1:8880/mailconfig/[email protected] proxyname: (null) proxyport: 0
[Wed Mar 11 11:14:31.808801 2020] [proxy_fcgi:debug] [pid 6902] mod_proxy_fcgi.c(916): [client *.*.*.*:53914] AH01077: declining URL http://127.0.0.1:8880/mailconfig/[email protected]
[Wed Mar 11 11:14:31.808812 2020] [proxy:debug] [pid 6902] proxy_util.c(2153): AH00942: HTTP: has acquired connection for (*)
[Wed Mar 11 11:14:31.808827 2020] [proxy:debug] [pid 6902] proxy_util.c(2206): [client *.*.*.*:53914] AH00944: connecting http://127.0.0.1:8880/mailconfig/[email protected] to 127.0.0.1:8880
[Wed Mar 11 11:14:31.808940 2020] [proxy:debug] [pid 6902] proxy_util.c(2415): [client *.*.*.*:53914] AH00947: connected /mailconfig/?emailaddress=[email protected] to 127.0.0.1:8880
[Wed Mar 11 11:14:31.809186 2020] [proxy:debug] [pid 6902] proxy_util.c(2881): AH02824: HTTP: connection established with 127.0.0.1:8880 (*)
[Wed Mar 11 11:14:31.809223 2020] [proxy:debug] [pid 6902] proxy_util.c(3048): AH00962: HTTP: connection complete to 127.0.0.1:8880 (127.0.0.1)
[Wed Mar 11 11:14:31.817283 2020] [proxy:debug] [pid 6902] proxy_util.c(2168): AH00943: *: has released connection for (*)
[Wed Mar 11 11:14:31.817301 2020] [proxy:debug] [pid 6902] proxy_util.c(2988): [remote 127.0.0.1:8880] AH02642: proxy: connection shutdown
[Wed Mar 11 11:14:31.817395 2020] [ssl:debug] [pid 6902] ssl_engine_io.c(1044): [client *.*.*.*:53914] AH02001: Connection closed to child 0 with standard shutdown (server ***.com:443)

 

[ducktales]

I have the same problem on our servers. Already removed the clientConfig lines from the panel.ini and tried 'plesk repair' on web, mail and installation.

When I send a test XML directly to the process at 127.0.0.1:8880 I get the same error 400 Bad Request:

test.xml

<?xml version="1.0" encoding="utf-8"?>
<Autodiscover xmlns="http://schemas.microsoft.com/exchange/autodiscover/mobilesync/requestschema/2006">
<Request>
<EMailAddress>[email protected]</EMailAddress>
<AcceptableResponseSchema>http://schemas.microsoft.com/exchange/autodiscover/mobilesync/responseschema/2006</AcceptableResponseSchema>
</Request>
</Autodiscover>

curl -X POST -k -v -d @test.xml http://127.0.0.1:8880/Autodiscover/Autodiscover.xml

* About to connect() to 127.0.0.1 port 8880 (#0)
*   Trying 127.0.0.1...
* Connected to 127.0.0.1 (127.0.0.1) port 8880 (#0)
> Post /Autodiscover/Autodiscover.xml HTTP/1.1
> User-Agent: curl/7.29.0
> Host: 127.0.0.1:8880
> Accept: */*
> Content-Length: 361
> Content-Type: application/x-www-form-urlencoded
>
* upload completely sent off: 361 out of 361 bytes
< HTTP/1.1 400 Bad Request
< Server: sw-cp-server
< Date: Sat, 28 Mar 2020 07:55:33 GMT
< Content-Type: text/html
< Content-Length: 915
< Connection: close
< ETag: "5e7b5aa8-393"
<
<!DOCTYPE html>
<html lang="en" dir="ltr">
<head>
    <meta charset="utf-8">
    <meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
    <title>400 Bad Request</title>
    <link rel="shortcut icon" href="/favicon.ico">
    <link rel="stylesheet" href="/ui-library/plesk-ui-library.css?3.5.0">
    <script type="text/javascript" src="/ui-library/plesk-ui-library.min.js?3.5.0"></script>
    <script type="text/javascript" src="/cp/javascript/vendors.js"></script>
    <script type="text/javascript" src="/cp/javascript/main.js"></script>
    <script type="text/javascript" src="/error_docs/uat.js?v1"></script>
<link href="/error_docs/app.css" rel="stylesheet"></head>
<body>
<div id="app"></div>
<script>window.__INITIAL_DATA__ = {"code":"400","message":"Bad Request"}</script>
<script type="text/javascript" src="/error_docs/app.js?89d33edd4917eae77067"></script></body>
</html>
* Closing connection 0

 

[ducktales]

I solved it on our end by going to Tools & Settings > Mail > Outgoing mail control > Server-wilde settings > Change settings. There I cleared the custom domain and checked Reconfigure DNS.

My guess is this somehow rebuilt the autodiscover config for the proxy, or there is a problem with custom domain configuration.