Issue High Apache CPU usage but my server uses Nginx the apache have been turn off for everysite

[domox]

Hi, i made the upgrade from Plesk 12.5 to Onyx and i have been notice that No matter if i'm not using Apache in any of my websites is taking a high uses of resources from my server. I turn off apache for every subscription in the server.

Another thing is Plesk has at /usr/local/PMM/tmp a very high disk space taken, is over 160GB.

I will let you a screenshot from my Server health:

 

[domox]

I deleted all the temp files at: rm -rf /usr/local/PMM/tmp/* and the plesk is working perfectly but i was a little bit afraid that this harm something.

So, I'm still having the Apache issue.

 

[domox]

I forgot to say that my environment has CentOS 7.2

 

[domox]

@mizar

can you help me with this, i don't know why Apache is taking so many resources from my server if im not using it. I turn off Apache for all the sites im using nginx.

 

[IgorG]

Apache still serve all your webmails, for example, even it is disables for all your sites.

 

[domox]

Apache still serve all your webmails, for example, even it is disables for all your sites.

I understand that @IgorG buy I'm not using any email account and other thing is that when the traffic of the server is getting hight the Apache take resurces in the same way so, something has plesk bad in the configuration .

 

[Xavier12]

Hey guys,

I am having the same exact issue as well. Apache isn't used for any websites, but its spiking randomly ever since the Plesk Onyx update. This was never an issue on 12.5 prior to the Plesk Onyx update. I am using Cloudlinux 6.8. It can't be Emails or anything else since no major changes have occurred.

Please advise, thanks.

 

[domox]

Hey guys,

I am having the same exact issue as well. Apache isn't used for any websites, but its spiking randomly ever since the Plesk Onyx update. This was never an issue on 12.5 prior to the Plesk Onyx update. I am using Cloudlinux 6.8. It can't be Emails or anything else since no major changes have occurred.

Please advise, thanks.

That's exactly my problem and still persist, after i upgrade to the Onyx is this problem with the 12.5 i didn't have it. I rechecked all my configuration files and i did not find anything.

 

[R3DC0D3]

Hi there,

I got same issue still awating for right answer...

Please note: All email services disabled.

 

[Bitpalast]

Run
# ps aux | grep httpd
and identify the httpd processes with the highest CPU load.

Then run
# strace -p PID
with the PID of these processes. From the strace output identify the website that is causing the load.

Then go into /var/www/vhosts/SUBSCRIPTION/logs/access_log and /error_log to see what scripts are being run by that website that is causing the high load.

 

[R3DC0D3]

Many thanks @Peter Debik

 

[domox]

I did that but the trace do not shows me any subscription taking resurces it shows me this:
socket(PF_LOCAL, SOCK_DGRAM|SOCK_CLOEXEC, 0) = 77
sendmsg(77, {msg_name(21)={sa_family=AF_LOCAL, sun_path="/run/systemd/notify"}, msg_iov(1)=[{"READY=1\nSTATUS=Total requests: 4"..., 88}], msg_controllen=0, msg_flags=0}, MSG_NOSIGNAL) = 88
close(77) = 0
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
socket(PF_LOCAL, SOCK_DGRAM|SOCK_CLOEXEC, 0) = 77
sendmsg(77, {msg_name(21)={sa_family=AF_LOCAL, sun_path="/run/systemd/notify"}, msg_iov(1)=[{"READY=1\nSTATUS=Total requests: 4"..., 88}], msg_controllen=0, msg_flags=0}, MSG_NOSIGNAL) = 88
close(77) = 0
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}) = 0 (Timeout)
wait4(-1, 0x7ffe7d07a64c, WNOHANG|WSTOPPED, NULL) = 0
select(0, NULL, NULL, NULL, {1, 0}^CProcess 12133 detached

 

[domox]

the true is that Apache is not suppose to consume anything because is not in used by any domain or subscription. All the domains has the apache turn off and using nginx.

 

[Bitpalast]

the true is that Apache is not suppose to consume anything because is not in used by any domain or subscription. All the domains has the apache turn off and using nginx.

The fact is that is is consuming something and you will have to find out what it is doing to solve the issue.

It is possible that a domain name cannot be found in an strace output easily. Sometimes you need to repeat the command on different httpd instances. Sometimes it helps to restart httpd service and quickly afterwards identify the high load process and go into strace again on it.

 

[TimReeves]

See my thread https://talk.plesk.com/threads/shoc...ates-httpd-config-with-proxy-mode-off.340033/

The problem is that Plesk still generates Apache Configs even when "Proxy mode" is off. At the moment custom templates (for Apache, basically rip out most of the code) is the only solution.

 

[Xavier12]

See my thread https://talk.plesk.com/threads/shoc...ates-httpd-config-with-proxy-mode-off.340033/

The problem is that Plesk still generates Apache Configs even when "Proxy mode" is off. At the moment custom templates (for Apache, basically rip out most of the code) is the only solution.

Hi Tim,

Thanks for this. I kind of figured this could have been the issue all along.. Finally, an actual lead to solving non-stop apache errors for high cpu usage when only using nginx. I have been in contact back and forth directly with Plesk premium support for over a month and they could not find the solution.