litespeed randomly not responding for several minutes

[tommynsong]

We have a odd behavior where litespeed would randomly not responding for 3 - 5 minutes daily. We have enabled debug logs with highest settings and there is gap in the log during that event. Core dump is enabled as well, but there isn't any core file created.
So basically, there is a problem somewhere, but we can't figure it out due to no log and/or core dump.

Litespeed 4.1.4 enterprise x64

Any advise on what is the next course of action.

Regards,
Tommy S.

 

[webizen]

could be your network problem that prevent from accessing during the time in question.

you may setup a cronjob (every min) to request a resource from the server locally (sth like 'wget localhost/a_test_resource' and see if the problem persists.

 

[tommynsong]

Hi webizen,

That is not the case, I was actively login to the server during that time. We have Nagios monitoring the site and it return socket timeout during that period.

Thanks,
Tommy S.

 

[mistwang]

check ip_conntrack error from "dmesg" output.
check server load and "top" output when it happens, check I/O wait, and load.
if you can connect to server from local host, but failed from Nagios, it indicates that it is a network issue, specifically for the HTTP port.