• Please be aware: Kaspersky Anti-Virus has been deprecated
    With the upgrade to Plesk Obsidian 18.0.64, "Kaspersky Anti-Virus for Servers" will be automatically removed from the servers it is installed on. We recommend that you migrate to Sophos Anti-Virus for Servers.
  • The Horde webmail has been deprecated. Its complete removal is scheduled for April 2025. For details and recommended actions, see the Feature and Deprecation Plan.
  • We’re working on enhancing the Monitoring feature in Plesk, and we could really use your expertise! If you’re open to sharing your experiences with server and website monitoring or providing feedback, we’d love to have a one-hour online meeting with you.

Panel 10.4.4 stopped working, sw-cp-server doesn't run

Tony_C

New Pleskian
Hi, I've got a problem and I've been trying to work it out for a few days.

I run Plesk 10.4.4, CentOS5 - 2.6.18-028stab095.1 #1 SMP Mon Oct 24 20:15:15 MSD 2011 x86_64 x86_64 x86_64 GNU/Linux

At the weekend I was making a number of small changes to configuration files - stuff that didn't seem to be covered by Plesk. They *seemed* harmless enough.

For example, I changed some logging parameters, installed logwatch, changed httpd.conf to use my email as the admin email.

Everything seemed fine - I was restarting various services as I went along, and the panel was still working. I really was just making what I thought were small changes. I know my way around.

I then chose to get a StartSSL certificate. I followed instructions from here http://stevejenkins.com/blog/2011/09/how-to-use-a-free-startssl-certificate-in-postfix-for-ssltls/

I got as far as making a directory and running openssl req -new -newkey rsa:32768 -nodes -keyout hostname.domain.key -out hostname.domain.csr (with my domain of course).

As you'd expect with a 32768 bit key, this took a few hours.

While it was generating the key, the panel stopped working. Of course, that is a coincidence - but this is why it's hard - I'd restarted various things and the panel had still worked; I have no idea why it would stop working when I was leaving the machine alone to generate the key.

When I go to mydomain:8443, first I get a "the connection was reset", and if I reload the page, I get a "Firefox can't establish a connection to the server".

I've checked my core access_log, error_log, httpsd_access_log, messages, secure, every log file I can think of.

The only damaging thing I can think of is that I got annoyed at the little bug that generates new *.include files for every virtual host every day, and I deleted the oldest in each directory. Seems harmless cos I left 2 versions in each directory (I know psa cleans it up to the 5 most recent, but it was annoying me!)

So it's possible that the Plesk server is not getting configuration to listen on 8443. It would explain why I'm not getting any error logs - the thing starts up, doesn't find a config file, and maybe stops.

But, these files exist (yep, Plesk generates a new version of these as well every time any changes are made)

I can't think which of the changes I made might've caused this. I know, this is that typical user thing - but I honestly did try to be methodical! I did reverse changes, and I even include my name in most config files if I change things, so I can do a text search and reverse any changes that don't work.

For example, I noticed a lot of attacks against the default web server, so I tried putting in an .htaccess file. There was none there so I created one. It didn't have any effect, so I removed it. Those were the sort of changes I was making.

Please can someone guide me through how to diagnose this problem?

I have been trying to work it out through constant googling. It's hard - "panel not working" or even "plesk 8443 configuration" or any number of combinations aren't helping me. Every other forum posting has some other error with it - but my psa is starting fine, all my webservers are running fine, my email is fine, everything is lovely.

Therefore I'm fairly sure it's just one small thing, one simple config change, that I made. Generally, I reversed every change I could remember - even changing "serveradmin" back to its default setting.

I also tried 2 other things: 1) bootstrapper repair (but that is v10.13.4 and I'm on 10.4.4), which reported success, and 2) restoring configuration files from backup. I take a remote backup every hour - but I guess this is the downside of a great bit of software like Plesk: So much config is in databases and different parts of the file system.

Please let me know what you'd like me to post here and I'll happily comply :)
 
Just to add to this, what I've noticed is that I can execute "service sw-cp-serverd start", and I almost always get this:

Starting SWsoft control panels server... stale pidfile. [ OK ]

Then, a PS will give me this

502 10074 0.0 0.0 58996 1308 ? S 16:19 0:00 /usr/sbin/sw-cp-serverd -f /etc/sw-cp-server/config

If I then go to http://domain:8443, I get the "connection reset" message.

A PS shows that process 10074 is no longer executing.

Instead, I get this

psaadm 10076 1.7 0.3 183796 14576 ? Ss 16:19 0:00 /usr/bin/sw-engine-cgi -c /usr/local/psa/admin/conf/php.ini -d auto_prepend_file=auth.php3 -u p
psaadm 10110 0.0 0.1 183796 4500 ? S 16:19 0:00 \_ /usr/bin/sw-engine-cgi -c /usr/local/psa/admin/conf/php.ini -d auto_prepend_file=auth.php3
psaadm 10111 0.0 0.1 183796 4460 ? S 16:19 0:00 \_ /usr/bin/sw-engine-cgi -c /usr/local/psa/admin/conf/php.ini -d auto_prepend_file=auth.php3
psaadm 10112 0.0 0.1 183796 4460 ? S 16:19 0:00 \_ /usr/bin/sw-engine-cgi -c /usr/local/psa/admin/conf/php.ini -d auto_prepend_file=auth.php3

And when I try to load http://domain:8443, I get the second error message.

Those processes stay running - and if I re-run sw-cp-serverd I will get 4 more of these processes (which I would expect if it was running normally).
 
I've checked my core access_log, error_log, httpsd_access_log, messages, secure, every log file I can think of.
But you forget about main sw-cp-server log file - /var/log/sw-cp-server/error_log
 
Hi Igor - sorry, I wasn't clear. I did check that one, and this is what I get (those stops and starts were probably me trying it again and again to make it work :) :

2012-02-12 20:41:40: (log.c.75) server started
2012-02-12 20:52:39: (log.c.75) server started
2012-02-12 21:09:28: (log.c.75) server started
2012-02-12 21:09:47: (log.c.75) server started
2012-02-12 21:09:52: (log.c.135) server stopped
2012-02-12 21:09:58: (log.c.75) server started
2012-02-12 21:10:13: (log.c.75) server started
2012-02-12 21:30:35: (log.c.75) server started
2012-02-12 22:18:41: (log.c.75) server started
2012-02-12 22:31:40: (log.c.75) server started
2012-02-12 22:59:06: (log.c.75) server started
2012-02-12 22:59:06: (server.c.627) opening pid-file failed: /var/run/sw-cp-server.pid File exists
2012-02-13 07:01:59: (log.c.75) server started
2012-02-13 07:17:41: (log.c.75) server started
2012-02-13 14:43:41: (log.c.75) server started
2012-02-13 15:40:44: (log.c.75) server started
2012-02-13 15:45:36: (log.c.75) server started
2012-02-13 16:27:13: (log.c.75) server started
2012-02-13 16:28:41: (log.c.75) server started
2012-02-14 16:19:17: (log.c.75) server started
2012-02-14 17:51:39: (log.c.75) server started
2012-02-14 17:53:43: (log.c.75) server started

That's what makes it frustrating, because there's just no error there.
 
PROBLEM PARTIALLY SOLVED... but this is very odd.

I solved the problem, and I have no idea what to do about it.

As I mentioned, the panel stopped working when I was using openSSL to generate a huge key.

What I hadn't remembered was what also happened: StartSSL installs a certificate on your browser to verify your identity with it.

When this happened, the Plesk panel came up with a Firefox dialog box message saying "the page is asking you to identify yourself" using an SSL certificate. The only cert that was displayed was my personal cert from StartSSL in my browser. So I clicked that one.

The problem occurred EVER SINCE THEN.

I only found this because an update for Firefox just came out. I installed it and restarted - and then the Plesk panel page asked me to identify myself with a certificate again. This time, I clicked "cancel" - and the panel has loaded and is working properly.

But: The first thing you do when a web page won't work is, ask other people to try it. Because the problem might be on your computer.

So when the panel stopped working, I tried to access it using my mobile on a 3G network, and I tried it using different browsers.

Does this make any sense to anyone? It certainly is consistent with the fact that I've got nothing in my error logs. Have I somehow restricted access using certificates? That's something I never play with - I was generating the StartSSL certificate request as a way of seeing what the process looks like; I haven't changed any config files.

Is there any way I can get Plesk to recognise that certificate, or do I have to remove it? I don't want to keep clicking "cancel"...
 
Back
Top