CPU information on the container is different from the main node CPU info.
Node: CPU MHz: 2493.780 Container: CPU MHz: 623.445
It is absolutely normal. CPU information on the container differs from CPU information on the node in one parameter only: CPU MHz is 2493.780 on the node and 623.445 MHz on the container.
If we look closer at these numbers, we will see that 2493.780 = 623.445 x 4. That means the CPULIMIT value for the container is set to 25 (a percentage, i.e., 1/4th). You can check it by executing the following command on the Hardware Node:
# cat /etc/vz/conf/<Container_ID>.conf
You should see the CPULIMIT="25" string there.
CPULIMIT is one of the CPU resource parameters that you can configure and monitor for each container. It is a positive number indicating the CPU time, as a percentage, that the corresponding container is not allowed to exceed. For details refer to Administrator's Guide for PIM
CPULIMIT could be set to a container as described on the "Managing CPU Share" page of the Parallels Virtuozzo Containers for Linux User's Guide.
In rare cases when a hardware server has a bottle-neck due to heavy disk I/O activity caused by some containers, it is worth setting automatic CPULIMIT balancing -- just set a "
0" value for all containers. Remember to make a backup of the current config files (/etc/sysconfig/vz-scripts/*.conf) before applying changes:
# for vl in `vzlist -aHo veid` ; do vzctl set $vl --save --cpuunits 1000 --cpulimit 0 \ --burst_cpulimit 0 --burst_cpu_avg_usage 0 --cpus 0; done