Docker cpu限制的示例分析

  • Post author:
  • Post category:其他

了解Docker cpu限制的示例分析?这个问题可能是我们日常学习或工作经常见到的。希望通过这个问题能让你收获颇深。下面是小编给大家带来的参考内容,让我们一起来看看吧!
本文测试了,docker容器限制cpu资源使用的几个配置参数。分别使用top和dstat命令分析了资源占有情况。
package mainimport (    “flag”    “runtime”    “fmt”)func main() {    cpunum := flag.Int(“cpunum”, 0, “cpunum”)    flag.Parse()    fmt.Println(“cpunum:”, *cpunum)    runtime.GOMAXPROCS(*cpunum)    for i := 0; i < *cpunum – 1; i++ {        go func() {            for {            }        }()    }    for {    }}制作了一个测试cpu占用的镜像,镜像默认占满1个核心
FROM busyboxCOPY ./full_cpu  /full_cpuRUN chmod +x /full_cpuENTRYPOINT [“/full_cpu”, “-cpunum”]CMD [“1”]docker build -t fangfenghua/cpuuseset .docker push fangfenghua/cpuusesetdocker info…Default Runtime: runcSecurity Options: seccompKernel Version: 3.10.0-229.el7.x86_64Operating System: CentOS Linux 7 (Core)OSType: linuxArchitecture: x86_64CPUs: 4Total Memory: 993.3 MiBName: localhost.localdomainID: TU6M:E6WM:PZDN:ULJX:EWKS:      …docker run -it –rm=true  fangfenghua/cpuuseset [root@localhost src]# toptop – 07:23:52 up  1:23,  2 users,  load average: 0.61, 1.12, 1.04Tasks: 154 total,   3 running, 145 sleeping,   6 stopped,   0 zombie%Cpu(s): 18.0 us,  0.1 sy,  0.0 ni, 81.8 id,  0.0 wa,  0.0 hi,  0.1 si,  0.0 stKiB Mem :  1017144 total,   422120 free,   171676 used,   423348 buff/cacheKiB Swap:  1040380 total,  1040284 free,       96 used.   688188 avail Mem   PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND                                                                                  20196 root      20   0    3048    720    460 R 101.7  0.1   0:37.56 full_cpu                                                                                     1 root      20   0   41536   4028   2380 S   0.0  0.4   0:02.60 systemd                                                                                      2 root      20   0       0      0      0 S   0.0  0.0   0:00.04 kthreadd                                                                                     3 root      20   0       0      0      0 S   0.0  0.0   0:00.48 ksoftirqd/0                                                                                  5 root       0 -20       0      0      0 S   0.0  0.0   0:00.00 kworker/0:0H                                                                                 7 root      rt   0       0      0      0 S   0.0  0.0   0:00.69 migration/0   docker run -it –rm=true  fangfenghua/cpuuseset 4top – 07:27:17 up  1:27,  2 users,  load average: 2.41, 1.47, 1.18Tasks: 159 total,   3 running, 145 sleeping,  11 stopped,   0 zombie%Cpu(s): 99.6 us,  0.2 sy,  0.0 ni,  0.0 id,  0.0 wa,  0.0 hi,  0.3 si,  0.0 stKiB Mem :  1017144 total,   402508 free,   190908 used,   423728 buff/cacheKiB Swap:  1040380 total,  1040284 free,       96 used.   668608 avail Mem   PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND                                                                                  20935 root      20   0    3048    720    452 R 400.0  0.1   0:55.80 full_cpu                                                                                     1 root      20   0   41620   4088   2380 S   0.0  0.4   0:02.88 systemd                                                                                      2 root      20   0       0      0      0 S   0.0  0.0   0:00.04 kthreadd在linux 系统上,可以用来限制docker容器资源占用的参数有:
      –cpu-period int              Limit CPU CFS (Completely Fair Scheduler) period      –cpu-quota int               Limit CPU CFS (Completely Fair Scheduler) quota  -c, –cpu-shares int              CPU shares (relative weight)      –cpuset-cpus string          CPUs in which to allow execution (0-3, 0,1)docker提供了–cpu-period、–cpu-quota两个参数控制容器可以分配到的CPU时钟周期。–cpu-period是用来指定容器对CPU的使用要在多长时间内做一次重新分配,而–cpu-quota是用来指定在这个周期内,最多可以有多少时间用来跑这个容器。跟–cpu-shares不同的是这种配置是指定一个绝对值,而且没有弹性在里面,容器对CPU资源的使用绝对不会超过配置的值。
cpu-period和cpu-quota的单位为微秒(μs)。cpu-period的最小值为1000微秒,最大值为1秒(10^6 μs),默认值为0.1秒(100000 μs)。cpu-quota的值默认为-1,表示不做控制。
举个例子,如果容器进程需要每1秒使用单个CPU的0.2秒时间,可以将cpu-period设置为1000000(即1秒),cpu-quota设置为200000(0.2秒)。当然,在多核情况下,如果允许容器进程需要完全占用两个CPU,则可以将cpu-period设置为100000(即0.1秒),cpu-quota设置为200000(0.2秒)。
使用本文制作的容器镜像来测试,cpu-period和cpu-quota两个参数吧。
在本文使用的4核心系统中,如果希望cpuusetest占满两个核心,在如何配置呢?从上文的分析中可以看到,如果将cpu-period设置为100000,那么期望占满4个核心,则需要将cpu-quota设置为4*100000,期望占满一个核心则可设置为2*100000。下面就测试一下吧:
docker run –name cpuuse -d –cpu-period=100000 –cpu-quota=200000 fangfenghua/cpuusetest 4top – 07:46:31 up  1:46,  2 users,  load average: 0.16, 0.21, 0.51Tasks: 168 total,   2 running, 142 sleeping,  24 stopped,   0 zombie%Cpu(s): 47.8 us,  0.1 sy,  0.0 ni, 51.9 id,  0.1 wa,  0.0 hi,  0.1 si,  0.0 stKiB Mem :  1017144 total,   364724 free,   227816 used,   424604 buff/cacheKiB Swap:  1040380 total,  1040284 free,       96 used.   631052 avail Mem   PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND                                                                                  21766 root      20   0    3048    724    464 R 193.3  0.1   1:00.37 full_cpu                                                                                     1 root      20   0   41620   4088   2380 S   0.0  0.4   0:03.13 systemd                                                                                      2 root      20   0       0      0      0 S   0.0  0.0   0:00.05 kthreadd                                                                                     3 root      20   0       0      0      0 S   0.0  0.0   0:00.52 ksoftirtop – 07:47:17 up  1:47,  2 users,  load average: 0.47, 0.26, 0.51Tasks: 172 total,   3 running, 144 sleeping,  25 stopped,   0 zombie%Cpu(s): 99.6 us,  0.1 sy,  0.0 ni,  0.3 id,  0.0 wa,  0.0 hi,  0.1 si,  0.0 stKiB Mem :  1017144 total,   358760 free,   233292 used,   425092 buff/cacheKiB Swap:  1040380 total,  1040284 free,       96 used.   625180 avail Mem docker run –name cpuuse -d –cpu-period=100000 –cpu-quota=400000 fangfenghua/cpuusetest 4  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND                                                                                  21976 root      20   0    3048    724    456 R 398.3  0.1   0:16.81 full_cpu                                                                                 21297 root      20   0       0      0      0 S   0.3  0.0   0:00.08 kworker/0:2                                                                                  1 root      20   0   41620   4088   2380 S   0.0  0.4   0:03.19 systemd                                                                                      2 root      20   0       0      0      0 S   0.0  0.0   0:00.05 kthreadd使用上述两个参数可以,设置cpu的精确控制。还有一个参数cpu-share,是个相对值。假如设置A容器cpu-share为1536,设置B容器为512。那么,在容器B启动前,cpu占用情况为是什么呢?
top – 07:56:10 up  1:56,  2 users,  load average: 0.75, 0.36, 0.50Tasks: 153 total,   3 running, 140 sleeping,  10 stopped,   0 zombie%Cpu(s): 99.7 us,  0.1 sy,  0.0 ni,  0.0 id,  0.0 wa,  0.0 hi,  0.3 si,  0.0 stKiB Mem :  1017144 total,   436300 free,   155616 used,   425228 buff/cacheKiB Swap:  1040380 total,  1040284 free,       96 used.   703544 avail Mem   PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND                                                                                  22216 root      20   0    3048    720    456 R 399.3  0.1   0:55.03 full_cpu                                                                                     1 root      20   0   41620   4088   2380 S   0.0  0.4   0:03.29 systemd                                                                                      2 root      20   0       0      0      0 S   0.0  0.0   0:00.05 kthreadd                                                                                     3 root      20   0       0      0      0 S   0.0  0.0   0:00.54 ksoftirqd/0启动容器B:
top – 07:57:09 up  1:57,  2 users,  load average: 3.55, 1.16, 0.76Tasks: 162 total,   4 running, 148 sleeping,  10 stopped,   0 zombie%Cpu(s): 99.6 us,  0.2 sy,  0.0 ni,  0.0 id,  0.0 wa,  0.0 hi,  0.3 si,  0.0 stKiB Mem :  1017144 total,   428772 free,   158304 used,   430068 buff/cacheKiB Swap:  1040380 total,  1040284 free,       96 used.   700444 avail Mem   PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND                                                                                  22216 root      20   0    3048    720    456 R 305.7  0.1   4:40.78 full_cpu                                                                                 22336 root      20   0    3048    720    460 R  95.3  0.1   0:09.02 full_cpu                                                                                     1 root      20   0   41620   4088   2380 S   0.0  0.4   0:03.31 systemd从上述测试结果不难看出。设置相对数值时,容器B启动之前,容器A仍然占满了cpu,而容器B启动后则,容器占3/4,容器B占1/4。

云专线拥有端到端优化的协议与灵活的架构,并支持从第四层至第七层的各种网络协议,可支持网站、应用程序、在线平台等多种加速需求。
还有一个参数cpu-sets,指定容器使用的核心。使用上述测试容器测试,指定容器使用0,3核心:
docker run –name cpuuse -d –cpuset-cpus=0,3  fangfenghua/cpuusetest 40,3核心占用率:
[root@localhost src]# dstat -c -C 0,3——-cpu0-usage————–cpu3-usage——usr sys idl wai hiq siq:usr sys idl wai hiq siq 25   9  66   0   0   0: 12   1  87   0   0   0100   0   0   0   0   0:100   0   0   0   0   0 99   0   0   0   0   1:100   0   0   0   0   0 99   1   0   0   0   0: 99   1   0   0   0   0100   0   0   0   0   0:100   0   0   0   0   0100   0   0   0   0   0:100   0   0   0   0   01,2核心占用率:
[root@localhost src]# dstat -c -C 1,2——-cpu1-usage————–cpu2-usage——usr sys idl wai hiq siq:usr sys idl wai hiq siq 21   8  71   0   0   0: 10   1  89   0   0   0  0   0 100   0   0   0:  0   0 100   0   0   0  0   0 100   0   0   0:  0   0 100   0   0   0  0   0 100   0   0   0:  0   0 100   0   0   0  0   0 100   0   0   0:  0   0 100   0   0   0感谢各位的阅读!看完上述内容,你们对Docker cpu限制的示例分析大概了解了吗?希望文章内容对大家有所帮助。