Linux cpu load learning notes and load learning notes

Linux cpu load learning notes and load learning notes
Measure the current CPU workload by Load in linux

Load Average is the Average Load within a period of time (1 minute, 5 minutes, 15 minutes.

[root@CNC-BJ-5-3N1 ~]# w20:01:55 up 76 days, 8:20, 6 users, load average: 1.30, 1.48, 1.69

Meaning of the load average Value

• Single-core processor

Assume that our system is single-CPU, single-core, which is a one-way road, and the CPU task is compared to a car. When there are not many cars, load <1; when the car occupies the whole road, load = 1; when the road is full and the road is full of cars, load> 1

• Multi-core processor

We often find that the server Load> 1 is still running well because the server is a Multi-core processor ). If our server CPU is 2 cores, it means we have 2 roads. When we Load = 2, all roads are full of vehicles.

View the number of CPU Cores

# View CPU core grep 'model name'/proc/cpuinfo | wc-l # CPU information:/proc/cpuinfo # memory information:/proc/meminfo

Be cautious about the Load average Value

• 0.7 <load <1: This is a good situation. If you come in with more cars, you can still cope with the road.
• Load = 1: Your road is about to be congested, and there are no additional tasks for more resources. Check out what happened.
• Load> 5: very congested. Our road is very busy and every car cannot run very quickly.

Which of the following are the three Load values?

Generally, let's take a look at the 15-minute load. If the load is high, let's take a look at the load for 1 minute and 5 minutes to see if there is a downward trend. If the 1-minute load value is greater than 1, we don't have to worry about it, but if the load exceeds 1 in 15 minutes, we need to hurry and see what happened. Therefore, we need to view the three values based on the actual situation.