Resolved Plesk Panel 17.5 | 400 Bad Request The plain HTTP request was sent to HTTPS port

[WebHostingAce]

Hi,

Since upgrade to Plesk 17.5 from 17.0,

I get this error,

400 Bad Request
The plain HTTP request was sent to HTTPS port
nginx

Servers running CentOS 7 and 6.9

Plesk login no loger redirect to HTTPS.

Possibly a Bug.

 

[UFHH01]

Hi AusWeb,

pls. use the HTTPS version, instead of HTTP, when you try to login into your Plesk Control Panel:

Example ( recommended URLs ):

https://servername.hosting-provider-FQDN.com:8443
https://servername.YOUR-DOMAIN.com:8443

If you desire to use the HTTP port of your Plesk Control Panel, then you would choose the URLs:

http://servername.hosting-provider-FQDN.com:8880
http://servername.YOUR-DOMAIN.com:8880

 

[WebHostingAce]

Thank you for your input @UFHH01

It used to redirect http://servername.hosting-provider-FQDN.com:8443 to https://servername.hosting-provider-FQDN.com:8443

Since upgrade to 17.5 it doesn't work.

Also any domain on the server http://anydomain.tld:8443 no longer redirect to https://servername.hosting-provider-FQDN.com:8443

 

[UFHH01]

Hi AusWeb,

pls. check the "sw-cp-server" configuration files and the corresponding logs at "/var/log/sw-cp-server" for further investigations.

 

[Kingsley]

I am having the same issue 17.5 also

 

[Kingsley]

working again for now

 

[Kingsley]

Hi AusWeb,

pls. use the HTTPS version, instead of HTTP, when you try to login into your Plesk Control Panel:

Example ( recommended URLs ):

https://servername.hosting-provider-FQDN.com:8443
https://servername.YOUR-DOMAIN.com:8443

If you desire to use the HTTP port of your Plesk Control Panel, then you would choose the URLs:

http://servername.hosting-provider-FQDN.com:8880
http://servername.YOUR-DOMAIN.com:8880

Why is HTTP not redirecting to HTTPS?

 

[WebHostingAce]

Probably a small bug in Plesk Panel 17.5.

 

[Kingsley]

Probably a small bug in Plesk Panel 17.5.

I guess, seems it has happened in 11.5 before

 

[UFHH01]

Hi Kingsley,
Hi AusWeb,

before you speculate further, you really should consider to post the corresponding configuration files and log - files, so that people willing to help you are actually able to investigate the issue/error/problem.

 

[CharisM]

Hi, i am having the same problem on 3 ! different server after upgrading to the latest version. no http to https redirection. Plesk was upgraded from 12.5 and i never had this before nor changed any config files. Error, access, sw-engine logs are empty. Redirecting was working properly before. vServers are running on ubuntu 14.04. I tried to fix this bug with the tipps for 11.5 plesk version but that did not work. would be great to fix that! thanks and best, c--

 

[Laughing Steam]

Same issue here. Not sure what's going on, but here's the error log in /var/log/sw-cp-server/

2017/04/25 11:20:07 [error] 10193#0: *989 empty URI in redirect to named location "@err" while reading client request line, client: REMOVED, server: , request: "

and from the sw-engine log:

[25-Apr-2017 18:03:28] NOTICE: Terminating ...
[25-Apr-2017 18:03:28] NOTICE: exiting, bye-bye!
[25-Apr-2017 18:03:28] NOTICE: fpm is running, pid 14301
[25-Apr-2017 18:03:28] NOTICE: ready to handle connections
[25-Apr-2017 18:03:28] NOTICE: systemd monitor interval set to 10000ms
[25-Apr-2017 18:04:26] NOTICE: Terminating ...
[25-Apr-2017 18:04:26] NOTICE: exiting, bye-bye!
[25-Apr-2017 18:04:26] NOTICE: fpm is running, pid 23740
[25-Apr-2017 18:04:26] NOTICE: ready to handle connections
[25-Apr-2017 18:04:26] NOTICE: systemd monitor interval set to 10000ms
[25-Apr-2017 18:18:38] NOTICE: Terminating ...
[25-Apr-2017 18:18:38] NOTICE: exiting, bye-bye!
[25-Apr-2017 16:18:38] NOTICE: fpm is running, pid 503
[25-Apr-2017 16:18:38] NOTICE: ready to handle connections
[25-Apr-2017 16:18:38] NOTICE: systemd monitor interval set to 10000ms
[25-Apr-2017 16:26:26] NOTICE: Terminating ...
[25-Apr-2017 16:26:26] NOTICE: exiting, bye-bye!
[25-Apr-2017 16:26:26] NOTICE: fpm is running, pid 7198
[25-Apr-2017 16:26:26] NOTICE: ready to handle connections
[25-Apr-2017 16:26:26] NOTICE: systemd monitor interval set to 10000ms
[25-Apr-2017 16:59:15] NOTICE: Terminating ...
[25-Apr-2017 16:59:15] NOTICE: exiting, bye-bye!
[25-Apr-2017 16:59:15] NOTICE: fpm is running, pid 22668
[25-Apr-2017 16:59:15] NOTICE: ready to handle connections
[25-Apr-2017 16:59:15] NOTICE: systemd monitor interval set to 10000ms

 

[WebHostingAce]

I think this issue is on every server running Plesk17.5. Im not sure if it is CentOS related since all my servers are running different versions of CentOS

Hopefully Plesk Team will fix this issue as soon as possible.

 

[Laughing Steam]

I'm running ‪Ubuntu 16.04.2 LTS, so it's with Plesk 17.5 I imagine as it seems to be happening to CentOS and Ubuntu. Hopefully it does get fixed soon!

 

[Bitpalast]

I can confirm the issue for Plesk 17.5 with latest MU on CentOS.

 

[UFHH01]

Hi @AlL,

I still wonder, why nobody opens a bug - report at

=> Home > Forum > Plesk Discussion > Reports​

... so that a Plesk - Team - Member ( for example @IgorG , @Aleksey Filatev , @snezymaev, @Viktor Vogel, ... ) is able to pass it over to the Plesk developpers.

 

[Bitpalast]

Personally I do not think it is a bug. It is a behavior, but not a bug. The response that Nginx delivers is correct. A request without SSL data is coming in on the SSL port. Why should Nginx forward this to an SSL encrypted connection?

 

[WebHostingAce]

@Peter Debik It was working before the upgrade. I really don't mind this minor issue. but some non-technical customers wouldn't like to see it. only my concern.

 

[UFHH01]

Hi Peter Debik,

hm...

Personally I do not think it is a bug. It is a behavior, but not a bug.

so you think, that the REDIRECT at the sw-cp-server - configuration ( as for example "error_page 497 https://servername.YOUR-DOMAIN.COM:$server_port$request_uri;" ) should be ignored? Pretty odd ( from my point of view ), if directives doesn't work the way, that they are expected to work as and the documentations ( nginx / Plesk ) are wrong then.

 

[Bitpalast]

Good point. It seems a bit like a philosophy issue. I understand that
a) The error_page directive does not work - and that is a bug, o.k.
b) It is a function that was there before, and that should not be taken away from users - 2 points for a bug, zero for correct behavior.
but
c) It can make sense to not auto-redirecting from non-SSL to SSL, maybe to make users aware that their initial request was unsecure? Give 1 point against the bug.
= result is +2 for a bug, +1 for not-a-bug.
o.k., agree that it should be considered as a bug.