Resolved Cant connect to Plesk Panel.

[Bitpalast]

That was not really the question.

The error.log is clearly showing that sw-cp-server cannot bind to the port, because the ports are already in use by another process. The only other process that binds to port 8443 is a running sw-cp-server. It can be a hanging process that cannot be stopped or started by a service command. For that reason the guide asks to kill such processes, then restart sw-cp-server.

Have you checked on "sw-cp..." in the process listed and killed the existing processes, then restarted sw-cp-server?

If so, what is the output of
# netstat -plnt | grep ':8443'
?
It should be something very similar to

tcp        0      0 0.0.0.0:8443            0.0.0.0:*               LISTEN      3231/sw-cp-server:
tcp6       0      0 :::8443                 :::*                    LISTEN      3231/sw-cp-server:

 

[Larsm]

tcp 0 0 0.0.0.0:8443 0.0.0.0:* LISTEN 11918/config
tcp6 0 0 :::8443 :::* LISTEN 11918/config

Gesendet von meinem Nexus 5X mit Tapatalk

 

[Bitpalast]

You're almost there. What "config" processes are in the # ps aux | grep config output?

 

[Larsm]

You're almost there. What "config" processes are in the # ps aux | grep config output?

When I do this with stopped sw-cp-server it shows this:

tomcat8 1089 0.2 3.7 2891788 150036 ? Sl 11:21 0:22 /usr/lib/jvm/default-java/bin/java -Djava.util.logging.config.file=/var/lib/tomcat8/conf/logging.properties -Djava.util.logging.manager=org.apache.juli.ClassLoaderLogManager -Djava.awt.headless=true -Xmx128m -XX:+UseConcMarkSweepGC -Djava.endorsed.dirs=/usr/share/tomcat8/endorsed -classpath /usr/share/tomcat8/bin/bootstrap.jar:/usr/share/tomcat8/bin/tomcat-juli.jar -Dcatalina.base=/var/lib/tomcat8 -Dcatalina.home=/usr/share/tomcat8 -Djava.io.tmpdir=/tmp/tomcat8-tomcat8-tmp org.apache.catalina.startup.Bootstrap start
root 3896 0.0 0.0 25364 2948 ? S 11:24 0:00 dovecot/config
root 3943 0.0 1.8 162904 75124 ? Ss 11:24 0:03 /usr/sbin/spamd -d --pidfile=/var/run/spamassassin.pid --create-prefs --daemonize --helper-home-dir=/var/qmail --max-children=5 --nouser-config --username=popuser --virtual-config-dir=/var/qmail/mailnames/%d/%l/.spamassassin

But
netstat -plnt | grep ':8443'

has no entry
Gesendet von meinem Nexus 5X mit Tapatalk

 

[Bitpalast]

I find it difficult to assist in this case, because with each answer you are introducing new facts, make changes that were not discussed or avoid a "yes"/"no" answer. For example the question was not what is in the list when sw-cp-server is stopped, it was simply "what's in the list".

Maybe the problem is not yet understood. The problem is that port 8443 is in use. For that reason sw-cp-server cannot bind to it, and for that reason the control panel cannot be reached. The software that is using port 8443 on a Plesk system is sw-cp-server. So when you start sw-cp-server it should show as /sw-cp-server in the netstat output. Obviously it does not, but instead a process named "config" is occupying the port. So you have to find out what process that is. As long as the "config" is occupying the port, sw-cp-server cannot use it. When you stop sw-cp-server and the process "config" is no longer visible in the netstat output, it is unlikely that you will be able to find it in the process list, too. I was asking for the process list output to find out WHICH config process, but as you have stopped sw-cp-server, therewith created a situation where the netstat port 8443 test returns zero output. This means that now the config process that we need to find cannot be identified.

One more try. Please re-create the situation where netstat returns the /config lines, for example by starting
# service sw-cp-server start

Then check again
# netstat -plnt | grep ':8443'
if it still as a "config" occupy port 8443.

Then run
# ps aux | grep config

There should now be more "config" processes in the ps output that can give a hint what is going on.

 

[Larsm]

I find it difficult to assist in this case, because with each answer you are introducing new facts, make changes that were not discussed or avoid a "yes"/"no" answer. For example the question was not what is in the list when sw-cp-server is stopped, it was simply "what's in the list".

Maybe the problem is not yet understood. The problem is that port 8443 is in use. For that reason sw-cp-server cannot bind to it, and for that reason the control panel cannot be reached. The software that is using port 8443 on a Plesk system is sw-cp-server. So when you start sw-cp-server it should show as /sw-cp-server in the netstat output. Obviously it does not, but instead a process named "config" is occupying the port. So you have to find out what process that is. As long as the "config" is occupying the port, sw-cp-server cannot use it. When you stop sw-cp-server and the process "config" is no longer visible in the netstat output, it is unlikely that you will be able to find it in the process list, too. I was asking for the process list output to find out WHICH config process, but as you have stopped sw-cp-server, therewith created a situation where the netstat port 8443 test returns zero output. This means that now the config process that we need to find cannot be identified.

One more try. Please re-create the situation where netstat returns the /config lines, for example by starting
# service sw-cp-server start

Then check again
# netstat -plnt | grep ':8443'
if it still as a "config" occupy port 8443.

Then run
# ps aux | grep config

There should now be more "config" processes in the ps output that can give a hint what is going on.

There are 3 processes when I execute ps aux | grep config

Tomecat, Spamassassin and dovecot

Gesendet von meinem Nexus 5X mit Tapatalk

 

[Larsm]

I find it difficult to assist in this case, because with each answer you are introducing new facts, make changes that were not discussed or avoid a "yes"/"no" answer. For example the question was not what is in the list when sw-cp-server is stopped, it was simply "what's in the list".

Maybe the problem is not yet understood. The problem is that port 8443 is in use. For that reason sw-cp-server cannot bind to it, and for that reason the control panel cannot be reached. The software that is using port 8443 on a Plesk system is sw-cp-server. So when you start sw-cp-server it should show as /sw-cp-server in the netstat output. Obviously it does not, but instead a process named "config" is occupying the port. So you have to find out what process that is. As long as the "config" is occupying the port, sw-cp-server cannot use it. When you stop sw-cp-server and the process "config" is no longer visible in the netstat output, it is unlikely that you will be able to find it in the process list, too. I was asking for the process list output to find out WHICH config process, but as you have stopped sw-cp-server, therewith created a situation where the netstat port 8443 test returns zero output. This means that now the config process that we need to find cannot be identified.

One more try. Please re-create the situation where netstat returns the /config lines, for example by starting
# service sw-cp-server start

Then check again
# netstat -plnt | grep ':8443'
if it still as a "config" occupy port 8443.

Then run
# ps aux | grep config

There should now be more "config" processes in the ps output that can give a hint what is going on.

When I search for the PID it says me that is the worker process

sw-cp-server: master process /usr/sbin/sw-cp-serverd -c /

Gesendet von meinem Nexus 5X mit Tapatalk

 

[Bitpalast]

When I search for the PID it says me that is the worker process
sw-cp-server: master process /usr/sbin/sw-cp-serverd -c /

Which PID are we talking about? Are you saying that when you run
# netstat -plnt | grep ':8443'
and get output that "config" is listening on port 8443 you take the PID from there and look into the ps list and that PID resolves to sw-cp-server master process?

 

[Larsm]

I do that: netstat -plnt | grep ':8443' grep the process from there

tcp 0 0 0.0.0.0:8443 0.0.0.0:* LISTEN 19325/config
tcp6 0 0 :::8443 :::* LISTEN 19325/config

I think the 19325 is the PID.

With htop --sort-key PID I open htop and go down to the list and the PID. It shows the sw-cp-server process

Show here

Gesendet von meinem Nexus 5X mit Tapatalk

 

[Bitpalast]

I got a hint that the "/config" is correct for Ubuntu I should have asked first, what OS you are on. My reference and test system is CentOS where it is "/sw-cp-server". So that entry is o.k.

We have to go back to where we started (previous page where I already pointed to the right article):
Plesk is down. Service sw-cp-server failed to start

This is the root cause, I am almost sure. You must be overlooking something when you go through the process list. There must be a hanging sw-cp-server process.

1) So next step (and hopefully the final one) is to stop sw-cp-server
# service sw-cp-server

2) And then check again
# ps aux | grep sw-cp
When you run this command, no sw-cp-server processes must be shown in the output. If you see sw-cp-server processes, kill the processes. Can you please execute these two steps and report the output of "ps aux | grep sw-cp" here?

 

[Larsm]

I got a hint that the "/config" is correct for Ubuntu I should have asked first, what OS you are on. My reference and test system is CentOS where it is "/sw-cp-server". So that entry is o.k.

We have to go back to where we started (previous page where I already pointed to the right article):
Plesk is down. Service sw-cp-server failed to start

This is the root cause, I am almost sure. You must be overlooking something when you go through the process list. There must be a hanging sw-cp-server process.

1) So next step (and hopefully the final one) is to stop sw-cp-server
# service sw-cp-server

2) And then check again
# ps aux | grep sw-cp
When you run this command, no sw-cp-server processes must be shown in the output. If you see sw-cp-server processes, kill the processes. Can you please execute these two steps and report the output of "ps aux | grep sw-cp" here?

Sorry only the grep command [emoji26]

root 25348 0.0 0.0 18344 1032 pts/1 S+ 20:41 0:00 grep --color=auto sw-cp

Gesendet von meinem Nexus 5X mit Tapatalk

 

[Bitpalast]

Make sure that no zombie processes exist:
# ps aux | grep Z

Look into the init process that an orphaned sw-cp-server process is not listed in that:
# ps -elf | awk '{if ($5 == 1){print $4" "$5" "$15}}'

If there are no processes with "sw-cp" substring, check the contents of the /etc/sw-cp-server directory, if a second configuration file exists, maybe a "backup" with a "listen 8443" directive, e.g.
# grep 8443 /etc/sw-cp-server/conf.d/*
It is correct when plesk.conf and ipv6_ports.inc have these lines. However, if there is a third result or another file with a listen directive on that port, it could maybe take precedence over the default Plesk settings. Are the files in /etc/sw-cp-server and subdirectories all the authentic, originally installed versions or did you make any changes to them?

In the vhost_ssl.conf file that you have been talking about on page (1) of this thread: Does that file contain a listen directive for port 8443?

 

[Larsm]

Make sure that no zombie processes exist:
# ps aux | grep Z

Look into the init process that an orphaned sw-cp-server process is not listed in that:
# ps -elf | awk '{if ($5 == 1){print $4" "$5" "$15}}'

If there are no processes with "sw-cp" substring, check the contents of the /etc/sw-cp-server directory, if a second configuration file exists, maybe a "backup" with a "listen 8443" directive, e.g.
# grep 8443 /etc/sw-cp-server/conf.d/*
It is correct when plesk.conf and ipv6_ports.inc have these lines. However, if there is a third result or another file with a listen directive on that port, it could maybe take precedence over the default Plesk settings. Are the files in /etc/sw-cp-server and subdirectories all the authentic, originally installed versions or did you make any changes to them?

In the vhost_ssl.conf file that you have been talking about on page (1) of this thread: Does that file contain a listen directive for port 8443?

I have 8 zombie processes

Gesendet von meinem Nexus 5X mit Tapatalk

 

[Larsm]

Make sure that no zombie processes exist:
# ps aux | grep Z

Look into the init process that an orphaned sw-cp-server process is not listed in that:
# ps -elf | awk '{if ($5 == 1){print $4" "$5" "$15}}'

If there are no processes with "sw-cp" substring, check the contents of the /etc/sw-cp-server directory, if a second configuration file exists, maybe a "backup" with a "listen 8443" directive, e.g.
# grep 8443 /etc/sw-cp-server/conf.d/*
It is correct when plesk.conf and ipv6_ports.inc have these lines. However, if there is a third result or another file with a listen directive on that port, it could maybe take precedence over the default Plesk settings. Are the files in /etc/sw-cp-server and subdirectories all the authentic, originally installed versions or did you make any changes to them?

In the vhost_ssl.conf file that you have been talking about on page (1) of this thread: Does that file contain a listen directive for port 8443?

I have made no changes to that directory and files

Gesendet von meinem Nexus 5X mit Tapatalk

 

[Larsm]

I think it's not an port problem. I change the port to 8444. The same issue. The plesk mobile app give me this error.

Gesendet von meinem Nexus 5X mit Tapatalk

 

[Bitpalast]

The Plesk Mobile App has nothing to do with this, and changing the port from default 8443 to another one will cause new issues. Of course the mobile app won't work with it.

Really sorry, but this is not the way problems can be solved. I suggest to hand this over to Plesk support or to reinstall the server. No more comments from me on it.

 

[Larsm]

It is fixed my the support of plesk.

"Hello,

The issue has been fixed.

The root cause of the issue was misconfiguration of sw-cp-server files, particularly in the "/etc/sw-cp-server/fastcgi_params" file.

I have set the correct setting in the "/etc/sw-cp-server/fastcgi_params" config from the test server.

I have made backup of the file /root/fastcgi_params.backup.
"

Gesendet von meinem Nexus 5X mit Tapatalk

 

[omexlu]

Little offtopic, the plesk support is such cases is for free?

 

[Larsm]

Little offtopic, the plesk support is such cases is for free?

I have an plesk license from my rootserver Provider. I didn't read something that I must pay for it.

It did an request over this site.

Help Center

Gesendet von meinem Nexus 5X mit Tapatalk