Resolved 421 Misdirected Request

[philglau]

Just here to also add the we were affected by this bug. I was able to restore our server from before it went down to this error, but shortly after restoring the working version of the VM hosting Plesk, it must be doing some sort of 'auto-update' that causes the problem again.

 

[Compy]

Cant get the mail working. Outlook on all clients are disconnected.

 

[TheBeatBali]

Same problem here. Applied the fix, which worked. The site is back on. But Plesk keeps on ramdomnly logging out when I'm on it, which doesn't give me a lot of confidence. Any ideas why does it continues logging out randomly?

 

[m0thman]

Asked earlier, anybody got an idea how many sites were affected globally. This kinda sucks.

 

[UHF]

In some harden environments without root access

echo -e "proxy_ssl_server_name on;\nproxy_ssl_name \$host;\nproxy_ssl_session_reuse off;" | sudo tee /etc/nginx/conf.d/fixssl.conf > /dev/null && sudo systemctl restart nginx

 

[King555]

Now, as the update has been released, should we revert the manual changes from the command (.../etc/nginx/conf.d/fixssl.conf...)?

 

[Igcorreia]

SOLUTION HERE: https://support.plesk.com/hc/en-us/...QkZzEkdDE3NTI4NDAwNDkkajEkbDAkaDE4OTc3Njc2OTg.

 

[mizar]

Hello!

> Now, as the update has been released, should we revert the manual changes from the command (.../etc/nginx/conf.d/fixssl.conf...)?

No, update actually do the same thing - it creates /etc/nginx/conf.d/fixssl.conf with the required directives.

 

[VNick]

https://docs.plesk.com/release-notes/obsidian/change-log/#plesk-18071-mu1 - just to be clear that the fix is available in Plesk 18.0.70.3 and 18.0.71.1.

 

[Wai Hong]

After the update or apply the hotfix. My WorPress Website cannot present normarly. It's able to show the text but the style.css and picture file has show “301 Moved Permenatly” . Any one also facing this problem ?
Example : asiaton.com.my

 

[The 8th]

Im not sure this issue is related, because I just woke up on it

nginx: [emerg] "proxy_ssl_server_name" directive is duplicate in
/etc/nginx/conf.d/z_fix_ssl_sni.conf:1
nginx: configuration file /etc/nginx/nginx.conf test failed
.................................................................. [ERROR]

 

[LRLD]

Im not sure this issue is related, because I just woke up on it

nginx: [emerg] "proxy_ssl_server_name" directive is duplicate in
/etc/nginx/conf.d/z_fix_ssl_sni.conf:1
nginx: configuration file /etc/nginx/nginx.conf test failed
.................................................................. [ERROR]

Yeah because the update to fix the issue didn't check if it was already fixed, it just added the directives regardless!
You'll need to delete the file you previously created.

 

[VNick]

In case you are using Plesk Premium Email extension and experience related issues with it after the hotfix, you may replace the /etc/nginx/conf.d/fixssl.conf content with:

proxy_ssl_server_name on;
proxy_ssl_session_reuse off;

map $host $webmail_host {
        ~^webmail\..+ $host;
        default webmail.$host;
}
map $request_uri $proxied_host {
       ~*^/autodiscover/autodiscover.xml $webmail_host;
       ~*^/(\.well-known/autoconfig/|)mail/config-v1.1.xml $webmail_host;
       ~*^/Microsoft-Server-ActiveSync $webmail_host;
       ~*^/iRony $webmail_host;
       default $host;
}
proxy_ssl_name $proxied_host;

and reload or restart the nginx service.

This is a workaround for a bug in PPE extension.

 

[BuzzLightBurger]

I looked at this code in another thread and it looks like it maps the xml/config files for each vhost domain as I'm using PPE but all my domains use the server's domain name for webmail and have no issues (aside from the repair I needed to do on the SQL following a kolab/roundcube upgrade the other day).

Would I be correct in saying that this code would be required if each domain used a specific webmail.vhost_domain.tld instead of the server wide domain name?