Page 1 of 1

steady httpd climb

Posted: Tue Jan 16, 2018 10:37 am
by chicjo01
Nagios Support,
We encountered a problem where the http service is running on the server, but the page would return an error as if httpd was not running. When I check the message logs and http error logs, I did not see anything that would indicate a problem or failure. Based on the perf graph below, you can see over time the http process seems to be getting used but not always released.

Do you know why this is happening and what additional information do you need to help with this?

Nagios XI: 5.4.2
Capture.PNG

Re: steady httpd climb

Posted: Tue Jan 16, 2018 2:27 pm
by npolovenko
Hello, @chicjo01. That gap on the graph indicates that Nagios check wasn't receiving any data at that point. It could be caused by a variety of reasons, like firewall interference, or Nagios could've lost connection to the Apache server. How often does this happen? Have you restarted the Apache server or Nagios server recently?

Re: steady httpd climb

Posted: Wed Jan 17, 2018 9:02 am
by chicjo01
The issue we are having doesn't have to do with the gap in the perf data. The problem is that the number of http processes keep a steady rise until it can no long server up the web interface of nagios.

Re: steady httpd climb

Posted: Wed Jan 17, 2018 4:04 pm
by npolovenko
@chicjo01, I see. I'd probably start by looking at the apache error and access log files. Would you be able to copy the whole /var/log/apache/ folder from your Linux system, put it in the zip file and upload here? Particularly, we'd like to see the log files from the 11-13th of January.

Re: steady httpd climb

Posted: Thu Jan 18, 2018 9:02 am
by chicjo01
I e-mailed [email protected], because the tarball was too large to upload.

Re: steady httpd climb

Posted: Thu Jan 18, 2018 11:44 am
by npolovenko
@chicjo01, They did forward me your email but there were no attachments. Can you send that email one more time to [email protected] ?Or you can upload the archive to a cloud storage and share a link with me in a PM.
Thanks

Re: steady httpd climb

Posted: Tue Jan 23, 2018 11:24 am
by chicjo01
My company has restricted e-mail so I can send out the log files and I have access to cloud storage. Do you have any other suggests for checking these logs? Do you have any keywords, I can check the logs?

Re: steady httpd climb

Posted: Tue Jan 23, 2018 11:26 am
by chicjo01
Can you increase the size of the file able to be uploaded from 1 mb to 10 mb?

Re: steady httpd climb

Posted: Tue Jan 23, 2018 3:01 pm
by npolovenko
@chicjo01, Can you take one acess_log, one ssl_error_log, and one error_log, put them in a zip file and send to me in a PM? These three files shouldn't be larger than 1mb. Also, I'd like to see the output of this command:

Code: Select all

netstat -anp | grep 80
If you're using a different port for httpd connections please change it in the command.
Also, seing the command and the command definition for this service check would be helpful.

PS: I'll put a request to our sysadmin to increase the 1mb threshold but it may take a while to get approved.
You could open a support ticket instead https://support.nagios.com/tickets/ They do allow a larger size file uploads.