Applicable to:
- Plesk Onyx for Linux
Symptoms
-
Apache log file ( CentOS based:
/var/log/httpd/error_logor Ubuntu based:/var/log/apache2/error.log) might show one of the following messages:CONFIG_TEXT: [notice] caught SIGTERM, shutting down
CONFIG_TEXT: [mpm_event:notice] [pid 665:tid 140662910175360] AH00492: caught SIGWINCH, shutting down gracefully
CONFIG_TEXT: [mpm_prefork:notice] [pid 12450] AH00169: caught SIGTERM, shutting down
CONFIG_TEXT: [error] 34053#0: *190501 upstream prematurely closed connection while reading response header from upstream, client: 203.0.113.101, server: example.com, request: "POST /wp-admin/admin-ajax.php HTTP/2.0", upstream: "https://203.0.113.2:7081/wp-admin/admin-ajax.php", host: "example.com", referrer: "https://example.com/"
-
System journal for Apache service contains multiple messages about Apache restart:
CONFIG_TEXT: Nov 27 08:13:00 servername systemd[1]: Stopping The Apache HTTP Server...
Nov 27 08:13:03 servername systemd[1]: Stopped The Apache HTTP Server.
Nov 27 08:13:03 servername systemd[1]: Starting The Apache HTTP Server...
Nov 27 08:13:05 servername systemd[1]: Started The Apache HTTP Server.
Nov 27 08:13:54 servername systemd[1]: Stopping The Apache HTTP Server...
Nov 27 08:13:57 servername systemd[1]: Stopped The Apache HTTP Server.
Nov 27 08:13:58 servername systemd[1]: Starting The Apache HTTP Server...
Nov 27 08:13:59 servername systemd[1]: Started The Apache HTTP Server. - Domains are unavailable with 503 error after issuing an SSL certificate or deleting a domain/subdomain
Cause
Plesk restarts Apache service to apply the new configuration when: a domain is created/removed, an SSL certificate has been issued/assigned, PHP Settings or Hosting settings for a domain have been changed.
Resolution
To avoid frequent Apache restarts, consider one of the following options:
-
Upgrade to Plesk Obsidian: By default, Apache service in Plesk Obsidian is reloaded instead to be restarted to avoid such downtimes
-
In Plesk Onyx, configure graceful restart for Apache
-
In Plesk Onyx, increase Apache restart interval by going to Tools & Settings > Apache Web Server > Set the Apache restart interval value to "300" or more (value in seconds):
Comments
9 comments
@Daniel Omine,
Great idea!
I can suggest you take part in our product improvement by referring to the following link: https://plesk.uservoice.com/forums/184549-feature-suggestions The top-ranked suggestions are likely to be included in the next versions of Plesk.
@Gjimi, I recommend increasing the Apache restart interval in Plesk > Tools & Settings > Apache Web Server and increase the Apache restart interval value to "300" or more.
If the issue persists we are gonna have to take a deeper look: https://support.plesk.com/hc/en-us/articles/213953025-How-to-get-support-directly-from-Plesk-
Which command is executed by Plesk for Restart or Reload Apache?
Anyhow, as follows would work well (Debian 9):
/usr/sbin/apache2ctl -k restart
and reload:
/usr/sbin/apache2ctl -k graceful
Please correct this in Plesk, it affects a lot of servers and that's not a condition.
(nginx is not installed)
@Daniel Omine,
Hello!
I understand your concern, however, Plesk update tool uses the native functionality of operating system package managers, such as yum or apt, to install updates. It does not verify packages, just installs them as if they were installed manually like as # yum update
More information is available in our Documentation.
Helps nothing. "graceful restart for Apache." unfortunately not, with graceful true then Plesk does nothing, no reload nothing, with false or without graceful is only stopped.
Apache Log:
body.xml:6: parser error : xmlParseEntityRef: no name
<param><value>1234%^&*</value></param>
^
[Thu Aug 08 20:49:21.185585 2019] [mpm_event:notice] [pid 14691:tid 140601467588672] AH00491: caught SIGTERM, shutting down
Plesk Version 17.8.11 / 17.8.xx Update XX, I think everyone is concerned
@Gimji, the commands you sent us are correct. However, if you are still receiving the errors please open a ticket with us so we can fix this issue: https://support.plesk.com/hc/en-us/articles/213953025-How-to-get-support-directly-from-Plesk-
@Julian, it is 300, everything is already tried.
I had this problem today after recent Plesk autoupdate restart the server.
After restart, all services were disabled. DNS, Mail, FTP, Http, SSH..
To solve this, i restarted the server manually, using the tools from the hosting server.
After restart, all services worked normally, except the http.
The strangest thing is, the http (apache) was running well.
Changing the database values doesn't caused any effect as this page teaching.
However, i noticed that the service "Reverse Proxy Server (nginx)" was not running.
I turned it on and all worked!!
I will never more trust on Plesk autoupdate. I disabled all.
I'm aware about this.
This server do not have any package installed manually.
All the things are installed by Plesk on default mode.
For me, this is a serious bug on Plesk.
Plesk should fix instead of pointing the problem to the server administrator.
"plesk uses the server native tools... "
It's the Plesk's responsability anyway.. Please, do not send excuses like this. Just try to found solution to avoid it on future. My feedback is here.
My suggestion is, before perform restart, the most obvious thing to do is check all the services running at that moment.
Create a list of applications running, restart and then, after the reboot, read the saved list to start the services one by one.
If some service from this list fail to wake up, send warning to the administrator.
That' all. Thank you for your time.
Please sign in to leave a comment.