all sites point to default vhost from time to time

[abdi]

So you have phpbb installed on your server? or can you install it? If phpbb does not redirect to port 7080 then sure its resolved ...

If you don't mind please let me know how it all goes ...

 

[abdi]

Just tested it and its not yet resolved.

 

[Achraf]

no disturbance

I test them with vbulletin forum and that's okay

 

[Achraf]

Its ok with PHPBB3
Its ok with Vbulletin
Its ok with 4images
Its ok with Wordpress
http://alrafi3.com/phpbb3/

 

[Bernard T]

Hello,

has the official fix been issued for this bug yet?

Maybe my investigation can toss more light into what could be wrong...

---------------------------------------------------------------
PRODUCT, VERSION, MICROUPDATE, OPERATING SYSTEM, ARCHITECTURE
Plesk Linux, 11.0.9 (upg. 10.4.4), #7, Debian 6.0, x64


PROBLEM DESCRIPTION
= = = = = = = = = =
"After some time all websites point to Default Page/Site or Nginx 404 error"

After several hours of frustrating manual investigation and log comparison,
I think I found a trace of circumstances when it comes to this issues.
When Nginx is active and we install an AppVault application in Plesk panel,
the Plesk system seems to "loose"/delete(?) all domain vhosts configuration
files, and then it goes to Nginx 404 Error (if Nginx is active) or to
Default Page/Site (if only Apache runs).

We had the following situation - first the Nginx was active. After the AppVault
App was installed, all sites got "Nginx - Error 404". We deactivated the Nginx
via Plesk, and the server was back on Apache only.
Several hours later came another AppVault installation and this brought us to
"Default page/site" issue. The only thing that helped then was command-line
vhosts regeneration.

I hope this helps developers to pinpoint and solve this issue!!


STEPS TO REPRODUCE
= = = = = = = = =
Situation (A) - with Nginx
-> upgrading already running and configured 10.4.4 system to 11.X (this step is important, I guess)
-> installing & activating Nginx via Plesk Installer (GUI)
-> installing an Application from AppVault (see ADDITIONAL INFO part for more clues from log)
==> all domains go "Nginx - Error 404"
-> deactivate Nginx in Plesk (Server Management > Services Management)

Situation (B) - without Nginx
-> AFTER the Nginx has been deactivated in Situation (A)
-> installing an Application from AppVault (see ADDITIONAL INFO part for more clues from log)
==> all domains go "Plesk/Apache Default site"


ACTUAL RESULT
= = = = = = =
Plesk (most probably) deletes all vhosts files and the :
(A) Nginx+Apache -> visiting any domain displays "Nginx 404 Error"
(B) Apache only -> visiting any domain displays "Plesk/Apache Default site"


EXPECTED RESULT
= = = = = = = =
to display domains normally


ANY ADDITIONAL INFORMATION
= = = = = = = = = = = = = =

- - - - - - -
Situation (A)
- - - - - - -
- Nginx: active
- Operation: installation of CloudFlare APS package

---> Plesk Actions Log
XX.YYY.WWW.ZZ plesk_clientA [2012-07-27 08:06:57] 'Install Site Application' ('Domain type': 'domain' => 'domain', 'Installation prefix': 'CDN' => 'CDN', 'Package name': 'CloudFlare for Plesk' => 'CloudFlare for Plesk')

---> /var/log/sw-cp-server/error.log
2012-07-27T08:06:42+00:00 ERR (3): Response from sappmng in stderr: TmpFile autoclean failed: No such file or directory

---> /var/log/apache2/error_log
[Fri Jul 27 08:07:04 2012] [notice] caught SIGTERM, shutting down

---> /var/log/nginx/error_log
2012/07/27 08:07:04 [error] 29745#0: *158927 connect() failed (111: Connection refused) while connecting to upstream, client: XX.YY.WWW.ZZ, server: some-domain.info, request: "GET / HTTP/1.1", upstream: "http ....
2012/07/27 08:07:05 [error] 29745#0: *158929 connect() failed (111: Connection refused) while connecting to upstream, client: XX.YY.WWW.ZZ, server: some-domain.info, request: "GET /favicon.ico HTTP/1.0", upstream: "http ....
......

((((( The Nginx is displaying ERROR 404 page )))))

=======> We tried to restart both, Nginx and Apache service, but none helped. Finally, disabling Nginx in Plesk Panel helped, but I suppose this acctually helped because it triggered the vhosts configs rebuilding process.
- - - - - - -


- - - - - - -
Situation (B)
- - - - - - -
- Nginx: deactivated in step (A)
- Plesk operation: installation of WordPress APS package

---> Plesk Actions Log
XX.YYY.WWW.ZZ plesk_clientB [2012-07-27 12:45:18] 'Install Site Application' ('Domain type': 'domain' => 'domain', 'Installation prefix': '' => '', 'Package name': 'Wordpress Instance' => 'Wordpress Instance')

---> /var/log/sw-cp-server/error_log
2012-07-27T12:45:17+00:00 ERR (3): Response from sappmng in stderr: TmpFile autoclean failed: No such file or directory
2012-07-27 12:52:59: (server.c.1533) unlink failed for: /var/run/sw-cp-server.pid 2 No such file or directory
2012-07-27 12:52:59: (server.c.1543) server stopped by UID = 0 PID = 19792
2012-07-27 12:54:11: (log.c.166) server started
2012-07-27 12:54:18: (mod_fastcgi.c.2746) FastCGI-stderr: PleskFatalException: Unable to connect to database: mysql_connect() [<a href='function.mysql-connect'>function.mysql-connect</a>]: No such file or directory
file: /opt/psa/admin/plib/common_func.php3
line: 153
code: 0
trace: #0 /opt/psa/admin/auto_prepend/auth.php3(112): psaerror('Unable to conne...')
#1 {main}

PHP Warning: mysql_errno() expects parameter 1 to be resource, boolean given in /opt/psa/admin/plib/common_func.php3 on line 209
PHP Warning: mysql_error() expects parameter 1 to be resource, boolean given in /opt/psa/admin/plib/common_func.php3 on line 222
PHP Warning: mysql_errno() expects parameter 1 to be resource, boolean given in /opt/psa/admin/plib/common_func.php3 on line 222
PHP Warning: Cannot modify header information - headers already sent by (output started at /opt/psa/admin/plib/PleskException.php:44) in /opt/psa/admin/plib/PleskException.php on line 27
...
...


---> /var/log/apache2/error.log
[Fri Jul 27 12:45:20 2012] [notice] caught SIGTERM, shutting down
[Fri Jul 27 12:45:22 2012] [error] [client XX.YYY.WWW.ZZZ] File does not exist: /var/www/vhosts/default/htdocs/some_folder_from_some_domain-com, referer: some_domain-com
[Fri Jul 27 12:45:22 2012] [error] [client XX.YYY.WWW.ZAQ] File does not exist: /var/www/vhosts/default/htdocs/other_folder_from_other_domain-com, referer: other_domain-com
[Fri Jul 27 12:45:23 2012] [error] [client XX.YYY.WWW.ZZZ] File does not exist: /var/www/vhosts/default/htdocs/some_folder_from_some_domain-com, referer: some_domain-com
[Fri Jul 27 12:45:23 2012] [error] [client XX.YYY.WWW.ZAQ] File does not exist: /var/www/vhosts/default/htdocs/other_folder_from_other_domain-com, referer: other_domain-com
....
1000x times - obviously the Apache can't find vhosts files, so it goes to /var/www/vhosts/default/htdocs/ (also, note the missing "TP" letters of "httpdocs" dir in this path)
....

((((( Apache is displaying Default Plesk page/site )))))

=======> BANDAID FIX: run /usr/local/psa/admin/bin/httpdmng --reconfigure-all

--------------------------------------------------------------

 

[rosskendall]

autoupdate caused problem

I have had a very similar problem.

Last night there was an autoupdate to Plesk 11 after that all the sites on the server went down (just showing the default page). The websites came back after resaving the hosting settings for each domain.

I have just had down time on that server with failed hard drive and had to reinstall, this extra bit of downtime was very unwelcome.

Please fix ASAP. (turned off autoupdates now, they can't be trusted if this is the result)

 

[IgorG]

Guys, developers can't reproduce this issue. Could you please send me in PM login credentials for your server? I will forward it to developers and they will check this issue directly on your server.

 

[Bernard T]

Guys, developers can't reproduce this issue. Could you please send me in PM login credentials for your server? I will forward it to developers and they will check this issue directly on your server.

OK, thanks for your reply. I will prepare the credentials for you.

In the meantime, as temporary precaution, we disabled the APS Application Installer option for all domains...

 

[abdi]

Guys, developers can't reproduce this issue. Could you please send me in PM login credentials for your server? I will forward it to developers and they will check this issue directly on your server.

I would love to, but since I can easily resolve that problem with /usr/local/psa/admin/sbin/httpdmng --reconfigure-all I can't risk that much down time as the developers check it out ...

NB: So far I have noticed that its quite common with Nginx enabled, if Apache fails to restart, if MySQL queries timeout pretty fast that /usr/local/psa/admin/sbin/httpdmng --reconfigure-all does not complete execution.

I think its something we might have control over other than a plesk problem ...

 

[Bernard T]

I think its something we might have control over other than a plesk problem ...

I'm not sure what you mean by that? ... I have several non-Plesk systems with Nginx and absolutely no such problems as here...

 

[Nikolay.]

/usr/local/psa/admin/sbin/httpdmng --reconfigure-all does not complete execution.

What do you mean by that? That reconfiguration takes so long that a MySQL connection is closed due to a timeout? How long does it take in your case?

 

[IgorG]

After investigation this issue directly on server of Bernard T we have released following article http://kb.parallels.com/en/114490

 

[abdi]

This is becoming so often with one of my servers showing plesk default page when opening webmail.domain.com especially with Nginx enabled.

 

[MislavO]

After investigation this issue directly on server of Bernard T we have released following article http://kb.parallels.com/en/114490

On the same server i started nginx recently and after some time again 404 nginx default pages. I fixed problem with just stoping nginx service (there was no need to reconfigure files, pages were working).


Nginx error log is full of errors, i don't want to copy everything, if needed, i can send again PM with the login credentials for server.

 

[Achraf]

After investigation this issue directly on server of Bernard T we have released following article http://kb.parallels.com/en/114490

the commande 'dpkg' not installed in centos.

 

[IgorG]

the commande 'dpkg' not installed in centos.

dpkg is Debian/Ubuntu utility for manage package system. Use the same utility for rpm-based OSes - rpm with corresponding options.

 

[MislavO]

On the same server i started nginx recently and after some time again 404 nginx default pages. I fixed problem with just stoping nginx service (there was no need to reconfigure files, pages were working).


Nginx error log is full of errors, i don't want to copy everything, if needed, i can send again PM with the login credentials for server.

I forgot to mention that we have new server where we installed clean Plesk 11 (no upgrade) and nginx is working there since first day and there are no problems so far, so i suppose this problem came along with the upgrade from 10.4.4 to 11.0.x. If needed, developers can check both servers if that will help with fixing this problem.

 

[Achraf]

after I use the command:
#rpm -e sw-sso

I can not get to a panel
https://mydomaine.net:8443/relay

Message: 500 - Internal Server Error

 

[Achraf]

after I use the command:
#rpm -e sw-sso

I can not get to a panel
https://mydomaine.net:8443/relay

Message: 500 - Internal Server Error

Problem solved with I desabled sso

#/usr/local/psa/bin/sso --disable
#/etc/init.d/psa restart

 

[abdi]

Sorry for the late reply, but yes that actually solves such a problem!