• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Issue Plesk is okay but sites are too slow or white page ?

SahinS

Basic Pleskian
Hello !

Everything was so beautiful until a couple of days ago. But some things go awry and now, sites are too slow.

I'm on Centos 6 & Plesk 17.5.3

I used to work PHP-FPM and Nginx. and I took some nginx errors and I removed Nginx and FPM from Plesk admin. Browser is trying to load any out web page but then all sites are white page. their error.log is empty.

What is your advice? How can watch the server well?

-----
some basic tryings:

# /usr/local/psa/bootstrapper/pp17.5.3-bootstrapper/bootstrapper.sh repair
.. is okay

# /usr/local/psa/bin/reconfigurator --autoconfigure
PHP:
mysql: [Warning] Using a password on the command line interface can be insecure.
Database: [<myip>]
Actual: [<myip>]
Removed IPs: [] []
Added IPs: [] []
Mapping: {}
Old not remapped: []
New not used: []
Nothing to do.

# /usr/local/psa/admin/sbin/httpdmng --reconfigure-all
.. is okay

maybe a clue:
onesite.com not our sites but its first nameserver is our server ip. ?

# /var/log/message
Code:
Jul  4 03:25:45 server named[23716]: client 46.4.113.114#56813: query (cache) 'oneSite.com/A/IN' denied
Jul  4 03:25:45 server named[23716]: client 46.4.113.114#40303: query (cache) 'oneSite.com/A/IN' denied
Jul  4 03:25:45 server named[23716]: client 46.4.113.114#57736: query (cache) 'oneSite.com/A/IN' denied
Jul  4 04:05:34 server named[23056]: client 74.125.181.2#42671: query (cache) 'www.oneSite.com/A/IN' denied
Jul  4 04:05:34 server named[23056]: client 74.125.47.14#42397: query (cache) 'www.oneSite.com/A/IN' denied
Jul  4 04:05:34 server named[23056]: client 74.125.47.132#33148: query (cache) 'www.oneSite.com/A/IN' denied

also,
I added this domain in prohibited domain list on Plesk admin.

but
all of services healt okay on plesk admin

:confused:
 
a more clue:

I realised an interesting thing that; if sites has ssl, they are working but others are not working.
 
Hi SahinS,

What is your advice? How can watch the server well?
The log - file investigations is always a good start, but you stated:
their error.log is empty.
... which leads to the question: Where did you look at?

The domain - specific webserver - log - files are located at: => /var/www/vhosts/system/(subdomain).YOUR-DOMAIN.COM/logs

#
/usr/local/psa/bootstrapper/pp17.5.3-bootstrapper/bootstrapper.sh repair
.. is okay
Pls. don't use this command anymore. You have now the "Plesk Repair Utility" with several options:

=> Plesk Repair Utility

Example command to automatically investigate and repair webserver - related issues/errors/problems:
Code:
plesk repair web -y -v


and I took some nginx errors and I removed Nginx and FPM from Plesk admin
Did you check your current domain - specific PHP AND webserver settings at:

=> HOME > Subscriptions > YOUR-DOMAIN.COM > PHP Settings ?​

In addition, we are not able to guess the ( possible ) issues/errors/problems from NGINX and or PHP-FPM. You really should consider to provide some corresponding log - entries... at the moment, we are a bit clueless as you are, as we are not able to investigate anything together with you. Pls. note some additional informations:


Sometimes, it is as well a good idea to change the log - level ( TEMPORARILY! ), to get more informations in Plesk - log - files:




maybe a clue:
onesite.com not our sites but its first nameserver is our server ip. ?
If you desire any help for DNS - investigations, pls. step out of anonymity and provide YOUR corresponding FQDN, so that people willing to help you have something to start with their investigations.

# /var/log/message
Not relevant here and the provided log - entries are "normal".


Sometimes it is as well a good idea to "reboot" your server ( you might have installed kernel - relevant updates/upgrades ).
 
Hi SahinS,
The log - file investigations is always a good start, but you stated:
.

Hi Dear @UFHH01
Again you ! First of all, thank for your reply! :)

Yes you are right. and now let me say you something step by step.

  1. Did you see my second post ? it can be essantional
    I realised an interesting thing that; if sites has ssl, they are working but others are not working. (It can be interested in RSA server certificate etc.? Because in the past, it done. But now, I am not sure, should I research and try this?)
  2. # plesk repair web -y -v
    all of them is okay there is just one error: The following domains have unregistered PHP handlers: ............. [ERROR] there are some domain name but we don't use them.

    I think, it isn't important. ?

  3. Yes I checked my PHP version , its normal, 5.6.3 , run PHP as FastCGI. as you know, I had removed Nginx and FPM
  4. About dns, there is a web site (it is not provided on our server. ) We don't know who is ? anyway, they saved their dns via our ip.
    and when I saw it log.file (query (cache) 'oneSite.com/A/IN' denied)

    I wanted to share with you, just this.

  5. I did reboot after these.
Sites are not perfect but they looks like it's getting better. after the plesk repair maybe ?

Thanks and Regards!
 
Hi SahinS,

Did you see my second post ?
I did, yes.

it can be essantional
It COULD be... but ( again ) sorry... I'm more the type who likes investigations more than guessings and a possible answer here can only be a guessing, as we have nothing to investigate ( no FQDN, no depending log - entries from your webserver/php... :(

there is just one error: The following domains have unregistered PHP handlers: ............. [ERROR] there are some domain name but we don't use them.
Pls. just consider to fix it by setting a valid PHP - handler for that domain. There is no reason to ignore the error - message.
I think, it isn't important. ?
You should be aware, that bots/visitors try to reach this domain and with constant issues/errors/problems ( caused by incorrect PHP - handler settings ), your webserver has to do some extra work, which you are able to solve by correcting the misconfiguration(s). ;)

About dns, there is a web site (it is not provided on our server. ) We don't know who is ? anyway, they saved their dns via our ip.
and when I saw it log.file (query (cache) 'oneSite.com/A/IN' denied)

I wanted to share with you, just this.
Yes, this can happen, but is no indication for the other described issues/errors/problems. Pls. consider to visit => Dns recurcion problem or attack? for further informations and solutions, how to solve this with the help of Fail2Ban. ;)

Sites are not perfect but they looks like it's getting better. after the plesk repair maybe ?
... and ( again ) sorry... but I can't give you a decent answer here, as we are not able to investigate your anonymous FQDN's. :(
 
Hi SahinS,

I did, yes.

Hi again,

I see. Yes, you are right!

About SSL RSA etc. I found an error in /var/www/vhosts/1domain.com/log/error_log . By the way, this site has no SSL certficate.

Code:
[Tue Jul 04 05:30:31 2017] [warn] RSA server certificate CommonName (CN) `Plesk' does NOT match server name!?

Now, should I try this ? Apache log: [warn] RSA server certificate CommonName (CN) `plesk' does NOT match server name!?
I am little anxious about trying this but if you say "yes", I will go on. :)

Thanks!
 
Hi SahinS,

[Tue Jul 04 05:30:31 2017] [warn] RSA server certificate CommonName (CN) `Plesk' does NOT match server name!?
You can safely ignore that WARNING ( no error! No other/further issues than constant messages about it in your logs! ), or if you desire to fix that, pls. visit the KB - article for further instructions. ;)
 
Hi SahinS,


You can safely ignore that WARNING ( no error! No other/further issues than constant messages about it in your logs! ), or if you desire to fix that, pls. visit the KB - article for further instructions. ;)

In that case, I forbeared. :)

I am thinking re-install Nginx and FPM , shall I install ? if so, are you ready for this ? :)
 
Hi SahinS,


Me? :D Sure... I'll be willing to help you at any time ( as well as the other forum users! ;) ).

Yes, of course as ever as you do. :)


Dear UFHH01,

  • I did re-install Nginx and FPM. At first, Nginx did not start and I tried to start as manuel but I saw this error:
    Code:
    Unable to start service: Unable to manage service by nginxmng: ('start', 'nginx'). 
    Error: [2017-07-06 07:16:52] ERR [util_exec] proc_close() failed ['/usr/local/psa/admin/bin/nginx_control' '--start'] with exit code [1] Error occured while sending feedback. 
    HTTP code returned: 502 Error occured while sending feedback. HTTP code returned: 502 Can not start proxy server:

    But it looks like was starting ? and then I used your sharing plesk repair code for web. Then, it was okay, too.

  • When I tried to restart Nginx on Plesk admin, it was successful.

  • But dynamic sites are giving on FPM application served by nginx mod

    404 Not Found nginx
  • If I try on FastCGI application served by Apache mod, they are working.
  • Last nginx log file like this:
    Code:
    2017/07/06 07:16:25 [emerg] 5268#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    
    2017/07/06 07:16:25 [emerg] 5132#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5132#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5212#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5212#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:24 [emerg] 4975#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:24 [emerg] 4975#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5268#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5268#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5132#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5132#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5212#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5212#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:24 [emerg] 4975#0: still could not bind()
    2017/07/06 07:16:25 [emerg] 5268#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5268#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5132#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5132#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5212#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5212#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5268#0: bind() to 5.135.242.107:80 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5268#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
    2017/07/06 07:16:25 [emerg] 5132#0: still could not bind()
    2017/07/06 07:16:25 [emerg] 5212#0: still could not bind()
    2017/07/06 07:16:25 [emerg] 5268#0: still could not bind()
  • # /usr/local/psa/admin/sbin/httpdmng --reconfigure-all
    was tried.
  • I saw there a post but I did not believe that. :confused:
    Plesk wordpress 404 Not Found nginx
What is your idea? o_O

Thanks,
Regards!
 
Hi SahinS,

2017/07/06 07:16:25 [emerg] 5268#0: bind() to 5.135.242.107:443 failed (98: Address already in use)
This indicates, that another process is using this port. Try to kill / stop such processes for example with:

HAMMER - method:

Code:
sudo fuser -k 80/tcp


Recommended - method:

Search for the actual process with for example:
Code:
netstat -ltnp | grep :80
and shutdown the processes with:
Code:
service SERVICE-NAME-FOUND-WITH-THE-PREVIOUS-COMMAND stop
... and restart nginx:
Code:
service nginx restart


# /usr/local/psa/admin/sbin/httpdmng --reconfigure-all
Pls. consider to use now:
Code:
plesk repair web -y -v
Your "httpdmng" is depricated by now and the Plesk Repair Utility investigates and repairs/re-creates new webserver - configuration files. At the end, a restart of your webserver related services will be done.

Additional informations:




404 Not Found nginx
This indicates missing nginx rewrites. Pls. be aware, that nginx can't read ".htaccess" - files and you have to use "Additional NGINX directives" ( => HOME > Subscriptions > YOUR-DOMAIN.COM > Apache and nginx Settings ) , so that nginx is able to serve the correct rewrites for your content. ;)
A usefull hint is to inspect your domain - specific log - files, in order to inform yourself about the used URL and the result from your webserver - it will mostly point you directly to your current issues/errors/problems.


I saw there a post but I did not believe that. :confused:
Plesk wordpress 404 Not Found nginx
Well the suggestions WAS correct... these days ... I wouldn't recommend the suggested changes anymore at all by now, though.

If you use wordpress, there are very an awfull lot of plugins, which use additional rewrites, mostly documented with ".htaccess" - rewrites, or they even implmented the automatic inserts of needed rewrites to your existing ".htaccess" - file(s). Often enough, the plugi authors don't care much about nginx - usage and forget to provide needed nginx directives. But ( !!! ) here are TWO good news:

  • In most cases, the "Additional nginx directive":
    Code:
        if (!-e $request_filename){
            rewrite ^(.*)$ /index.php break;
        }
    is mostly enough and if not, the plugin provider should have added additional informations to the plugin page. ;)
because I don't want to change every .htaccess for all sites as manuel :/
... is great... but pls. keep in mind, that this tool is not perfect - you can't fully rely on it. :)
 
Last edited by a moderator:
Hi SahinS,
..
  • In most cases, the "Additional nginx directive":
    Code:
        if (!-e $request_filename){
            rewrite ^(.*)$ /index.php break;
        }
    is mostly enough and if not, the plugin provider should have added additional informations to the plugin page. ;)
  • Your idea
... is great... but pls. keep in mind, that this tool is not perfect - you can't fully rely on it. :)

Hello!
I have an off-topic question. Do you advise me basic article about security and performance.

Thank you!
 
Back
Top