Linux CircleCI是如何实现的；“合并”；Docker cgroup名称空间？

我试图在CircleCI工作流中调试失败的测试作业，该工作流的配置类似于：

  integration_tests:
    docker:
      - image: my-group/our-custom-image:latest
      - image: postgres:9.6.11
        environment:
          POSTGRES_USER: user
          POSTGRES_PASSWORD: password
          POSTGRES_DB: db
    steps:
      - stuff && things

当我和宋承宪来到CircleCI应用程序告诉我的地方时，我发现自己陷入了一个奇怪的迷宫，迷宫中有许多扭曲的小名称空间，它们都是一样的。我运行了

ps awx

，可以看到来自两个docker容器的进程：

他们所做的似乎是以某种方式将两个docker容器的cgroup名称空间“合并”到第三个名称空间中，他们提供给我的shell位于该名称空间下。因为pid 7是从一个docker容器运行的，而pid 257是在

my group/our custom image:latest

容器中运行的应用程序

/proc

中的cgroup视图显示了正在进行的某种合并cgroup，它看起来像

root@90c93bcdd369:~# cat /proc/7/cgroup 
12:devices:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
11:blkio:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
10:memory:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
9:hugetlb:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
8:perf_event:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
7:net_cls,net_prio:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
6:rdma:/
5:cpu,cpuacct:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
4:cpuset:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
3:pids:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
2:freezer:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
1:name=systemd:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
0::/system.slice/containerd.service
root@90c93bcdd369:~# cat /proc/257/cgroup 
12:devices:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
11:blkio:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
10:memory:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
9:hugetlb:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
8:perf_event:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
7:net_cls,net_prio:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
6:rdma:/
5:cpu,cpuacct:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
4:cpuset:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
3:pids:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
2:freezer:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
1:name=systemd:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
0::/system.slice/containerd.service

是否有一个标准的Docker特性，或者一些cgroup特性被用来产生这种魔力？或者这是某种定制的、专有的CircleCI功能

root@90c93bcdd369:~# cat /proc/7/cgroup 
12:devices:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
11:blkio:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
10:memory:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
9:hugetlb:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
8:perf_event:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
7:net_cls,net_prio:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
6:rdma:/
5:cpu,cpuacct:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
4:cpuset:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
3:pids:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
2:freezer:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
1:name=systemd:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/d8fc5294708fd4cf91fa405d6462571e1dc56413b55a6b6e5790b8f158fee632
0::/system.slice/containerd.service
root@90c93bcdd369:~# cat /proc/257/cgroup 
12:devices:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
11:blkio:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
10:memory:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
9:hugetlb:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
8:perf_event:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
7:net_cls,net_prio:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
6:rdma:/
5:cpu,cpuacct:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
4:cpuset:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
3:pids:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
2:freezer:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
1:name=systemd:/docker/94376e7880579c6bde0622017594fcdb8d5767788bb4790c0f014db282198577/90c93bcdd3693a918adddf62939c5b31e86868864edabe7347a268149e797f43
0::/system.slice/containerd.service