If you can see this check that
Your system environment is as follows:
- Machine A, eth0 10.0.1.1/24, eth1 10.0.2.254/24.
- Machine B, eth0 10.0.2.1/24
- Machine C, eth0 10.0.2.2/24
- Machine B and C are connected to eth1 on A via a switch.
- Machine B runs apache (and its pages use Oracle on C to run) and telnet.
- Machine C runs Oracle.
- Machine A also runs apache.
You have noticed that the traffic on A-eth0 has doubled, as has A-eth1.
Strangely, C is using 100% of the CPU time. B has higher than usual loading
of apache, but has a CPU running at 80% idle. What could be the problem?
|A.||The network on A is too slow to cope with the traffic.
|B.||Apache on A is simply twice as popular as before.
|C.||B and C need to be moved to different physical networks.
|D.||The cgi programs on B have saturated Oracle on C with requests.
|E.||The network linking A,B, and C needs to be upgraded.
With high CPU loads detected, it is likely that the traffic loads only one of the factors to consider.
This is reasonable, but does not explain what is happening with B and C.
There is no evidence to support this.
This is certainly a possibility.
As the CPU load has reached 100%, more network resources probably will not help.