President-Anonymous Posted January 30, 2013 Report Share Posted January 30, 2013 I keep telling support that something is wrong and support keeps telling me everything is fine. You have 16 CPU Cores and you're trying to use 20+ Everything is not fine. Quote Link to comment Share on other sites More sharing options...
President-Anonymous Posted January 30, 2013 Author Report Share Posted January 30, 2013 My own personal usage -vs- the server stats above. Quote Link to comment Share on other sites More sharing options...
Tony Posted January 31, 2013 Report Share Posted January 31, 2013 Load average is measure of processes in queue it does directly correlated to the available CPU on a system. For example I'm logged into the system you're on it reports a load average is reporting as 10 with 50% of the CPU idle. If it was a direct correlation then this system would supposedly have 20 CPU's not the 16 it actually has. This used to be a pretty good indicator because there was very few reasons processes would be waiting in queue. This is no longer the case with using CloudLinux for example if a user is being CPU limited on old kernels if 5 processes of a single user were limited it would +5 the load average. Newer kernels of CloudLinux it instead will only +1 meaning if we had 20 users being CPU limited it would have a load average of 20 even if there was a lot of available CPU. The system you're on is running an older kernel with KSplice updates it may still have the older load average calculations as well. Basically load average while can be an indicator it should not be the final word. If your sites are fast no matter the load average there is nothing to worry about. Quote Link to comment Share on other sites More sharing options...
President-Anonymous Posted February 1, 2013 Author Report Share Posted February 1, 2013 Load average is measure of processes in queue it does directly correlated to the available CPU on a system. For example I'm logged into the system you're on it reports a load average is reporting as 10 with 50% of the CPU idle. If it was a direct correlation then this system would supposedly have 20 CPU's not the 16 it actually has. This used to be a pretty good indicator because there was very few reasons processes would be waiting in queue. This is no longer the case with using CloudLinux for example if a user is being CPU limited on old kernels if 5 processes of a single user were limited it would +5 the load average. Newer kernels of CloudLinux it instead will only +1 meaning if we had 20 users being CPU limited it would have a load average of 20 even if there was a lot of available CPU. The system you're on is running an older kernel with KSplice updates it may still have the older load average calculations as well. Basically load average while can be an indicator it should not be the final word. If your sites are fast no matter the load average there is nothing to worry about. Both my site and cPanel would not load when the CPU reading was high. This repeatedly kept happening. Deleted my site and thus was using ZERO resources and cPanel was still not loading smoothly. You have or had a system issue and your customer service basically said I was imagining it (just about called me crazy in the ticket). My review (and the review of many others) of your service will not be an imagination. Quote Link to comment Share on other sites More sharing options...
telogoreng Posted July 13, 2013 Report Share Posted July 13, 2013 is there any relation between cms that we ve used and overload cpu resources ? Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.