首先呢是发现了容器会重启的问题，然后查看日志报错
shitou@aishitou:~$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
ab357e46450f grafana/grafana:latest "/run.sh" 5 months ago Up 17 seconds 0.0.0.0:3000->3000/tcp, [::]:3000->3000/tcp grafana
553e54dcef0d prom/prometheus:latest "/bin/prometheus --c…" 5 months ago Up 17 seconds 0.0.0.0:9090->9090/tcp, [::]:9090->9090/tcp prometheus
eec3a45791bc google/cadvisor:latest "/usr/bin/cadvisor -…" 5 months ago Up 3 seconds 0.0.0.0:8080->8080/tcp, [::]:8080->8080/tcp cadvisor
查看
shitou@aishitou:~$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
ab357e46450f grafana/grafana:latest "/run.sh" 5 months ago Up 50 seconds 0.0.0.0:3000->3000/tcp, [::]:3000->3000/tcp grafana
553e54dcef0d prom/prometheus:latest "/bin/prometheus --c…" 5 months ago Up 51 seconds 0.0.0.0:9090->9090/tcp, [::]:9090->9090/tcp prometheus
eec3a45791bc google/cadvisor:latest "/usr/bin/cadvisor -…" 5 months ago Restarting (255) 4 seconds ago cadvisor
再次查看，
shitou@aishitou:~$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
ab357e46450f grafana/grafana:latest "/run.sh" 5 months ago Up 53 seconds 0.0.0.0:3000->3000/tcp, [::]:3000->3000/tcp grafana
553e54dcef0d prom/prometheus:latest "/bin/prometheus --c…" 5 months ago Up 54 seconds 0.0.0.0:9090->9090/tcp, [::]:9090->9090/tcp prometheus
eec3a45791bc google/cadvisor:latest "/usr/bin/cadvisor -…" 5 months ago Restarting (255) 7 seconds ago cadvisor
shitou@aishitou:~$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
ab357e46450f grafana/grafana:latest "/run.sh" 5 months ago Up 56 seconds 0.0.0.0:3000->3000/tcp, [::]:3000->3 000/tcp grafana
553e54dcef0d prom/prometheus:latest "/bin/prometheus --c…" 5 months ago Up 56 seconds 0.0.0.0:9090->9090/tcp, [::]:9090->9 090/tcp prometheus
eec3a45791bc google/cadvisor:latest "/usr/bin/cadvisor -…" 5 months ago Restarting (255) Less than a second ago cadvisor
报错内容就是如下
.0:8080->8080/tcp, [::]:8080->8080/tcp cadvisor
shitou@aishitou:~$ sudo docker logs eec3a45791bc
[sudo] password for shitou:
F0226 15:29:53.822922 1 cadvisor.go:146] Failed to create a Container Manager: mountpoint for cpu not found
F0226 15:29:58.289714 1 cadvisor.go:146] Failed to create a Container Manager: mountpoint for cpu not found
F0226 15:30:02.550374 1 cadvisor.go:146] Failed to create a Container Manager: mountpoint for cpu not found
F0226 15:30:06.154412 1 cadvisor.go:146] Failed to create a Container Manager: mountpoint for cpu not found
分析
从错误日志来看，cadvisor 启动失败的原因是无法找到 CPU 相关的挂载点（mountpoint for cpu not found）。这是因为 cadvisor 需要访问宿主机的 cgroup 目录来监控容器和系统资源，而默认情况下这些目录没有被正确挂载到容器中。
shitou@aishitou:~$ mount | grep cgroup | grep cpu
#目的是查看宿主机是否已挂载 CPU 相关的 cgroup 子系统（cpu 和 cpuacct 是 cgroup 中用于 CPU 资源控制的关键子系统）
#输出为空，说明当时系统中没有挂载 CPU 相关的 cgroup 子系统，这正是 cadvisor 报错的原因。
shitou@aishitou:~$ ls -ld /sys/fs/cgroup
dr-xr-xr-x 12 root root 0 Aug 14 07:30 /sys/fs/cgroup
#检查 cgroup 根目录是否存在（/sys/fs/cgroup 是 cgroup 子系统的默认挂载点）。
#输出显示目录存在（dr-xr-xr-x），权限正常，排除了目录不存在的问题。
shitou@aishitou:~$ sudo mkdir -p /sys/fs/cgroup/cpu /sys/fs/cgroup/cpuacct
[sudo] password for shitou:
shitou@aishitou:~$ sudo mount -t cgroup -o cpu cpu /sys/fs/cgroup/cpu #这里少写了
mount: /sys/fs/cgroup/cpu: cpu already mounted or mount point busy.
dmesg(1) may have more information after failed mount system call.
shitou@aishitou:~$ sudo mount -t cgroup -o cpuacct cpuacct /sys/fs/cgroup/cpuac ct
#手动创建 CPU 相关的 cgroup 子目录（因为之前检查发现没有挂载，可能是目录缺失导致）。
shitou@aishitou:~$ mount | grep cgroup | grep cpu
cpuacct on /sys/fs/cgroup/cpuacct type cgroup (rw,relatime,cpuacct)
#试图将 cpu 子系统挂载到 /sys/fs/cgroup/cpu。
--------------------------------------------------------------------------------------------------
shitou@aishitou:~$ sudo fuser -m /sys/fs/cgroup/cp
Specified filename /sys/fs/cgroup/cp does not exist.
报错 cpu already mounted or mount point busy说明该挂载点可能已被占用，或系统已通过其他方式挂载了 cpu 子系统。
--------------------------------------------------------------------------------------------------
上面错误，因为少写了一个U,哈哈cpu,
shitou@aishitou:~$ sudo fuser -m /sys/fs/cgroup/cpu
#没有输出任何内容，说明该目录当前没有被任何进程占用（fuser 命令会列出占用指定文件 / 目录的进程 ID，无输出即表示无进程占用）。
=========================================================================================
shitou@aishitou:~$ mount | grep cgroup | grep cpu
cpuacct on /sys/fs/cgroup/cpuacct type cgroup (rw,relatime,cpuacct)
#只显示了 cpuacct 子系统的挂载信息（/sys/fs/cgroup/cpuacct），而 cpu 子系统并未单独挂载。这与你系统使用 cgroup v2 的特性完全一致 ——v2 采用统一目录结构，所有子系统（包括 CPU 控制功能）都整合在 /sys/fs/cgroup 根目录下，无需也不支持单独挂载 cpu 子系统。
=========================================================================================
记住这个，因为我在后面改成了cgroup v1
=========================================================================================
shitou@aishitou:~$ sudo mount -t cgroup -o cpu cpu /sys/fs/cgroup/cpu
mount: /sys/fs/cgroup/cpu: cpu already mounted or mount point busy.
dmesg(1) may have more information after failed mount system call.
成功挂载了 cpuacct 子系统（后续 mount 命令显示其已挂载）。
shitou@aishitou:~$ mount | grep cgroup
cgroup2 on /sys/fs/cgroup type cgroup2 (rw,nosuid,nodev,noexec,relatime,nsdeleg ate,memory_recursiveprot)
cpuacct on /sys/fs/cgroup/cpuacct type cgroup (rw,relatime,cpuacct)
再次尝试挂载 cpu 子系统仍失败，最终通过 mount | grep cgroup 发现系统实际使用的是 cgroup v2（显示 cgroup2 on /sys/fs/cgroup），而你手动挂载的 cpuacct 是 cgroup v1 的子系统，两者可能不兼容。
shitou@aishitou:~$ docker start eec3a45791bc
eec3a45791bc
shitou@aishitou:~$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
ab357e46450f grafana/grafana:latest "/run.sh" 5 months ago Up 12 minutes 0.0.0.0:3000->3000/tcp, [::]:3000->3000/tcp grafana
shitou@aishitou:~$ docker start eec3a45791bc
eec3a45791bc
shitou@aishitou:~$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
ab357e46450f grafana/grafana:latest "/run.sh" 5 months ago Up 12 minutes 0.0.0.0:3000->3000/tcp, [::]:3000->3000/tcp grafana
553e54dcef0d prom/prometheus:latest "/bin/prometheus --c…" 5 months ago Up 12 minutes 0.0.0.0:9090->9090/tcp, [::]:9090->9090/tcp prometheus
eec3a45791bc google/cadvisor:latest "/usr/bin/cadvisor -…" 5 months ago Up 2 seconds 0.0.0.0:8080->8080/tcp, [::]:8080->8080/tcp cadvisor
shitou@aishitou:~$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
ab357e46450f grafana/grafana:latest "/run.sh" 5 months ago Up 13 minutes 0.0.0.0:3000->3000/tcp, [::]:3000->3 000/tcp grafana
553e54dcef0d prom/prometheus:latest "/bin/prometheus --c…" 5 months ago Up 13 minutes 0.0.0.0:9090->9090/tcp, [::]:9090->9 090/tcp prometheus
eec3a45791bc google/cadvisor:latest "/usr/bin/cadvisor -…" 5 months ago Restarting (255) Less than a second ago cadvisor
shitou@aishitou:~$ docker stop eec3a45791bc
eec3a45791bc
sudo docker logs eec3a45791bc
F0226 15:29:53.822922 1 cadvisor.go:146] Failed to create a Container Man ager: mountpoint for cpu not found
F0226 15:29:58.289714 1 cadvisor.go:146] Failed to create a Container Man ager: mountpoint for cpu not found
F0226 15:30:02.550374 1 cadvisor.go:146] Failed to create a Container Man ager: mountpoint for cpu not found
F0226 15:30:06.154412 1 cadvisor.go:146] Failed to create a Container Man ager: mountpoint for cpu not found
F0226 15:30:09.577217 1 cadvisor.go:146] Failed to create a Container Man ager: mountpoint for cpu not found
F0226 15:30:13.126774 1 cadvisor.go:146] Failed to create a Container Man ager: mountpoint for cpu not found
正在进一步解决 cgroup 相关的挂载问题（尤其是 cpu 子系统挂载失败），并尝试通过配置系统启动参数来调整 cgroup 版本（从 v2 切换到 v1，以更好地兼容 cadvisor）。
shitou@aishitou:~$ sudo vim /etc/fstab
shitou@aishitou:~$ sudo ls /etc/fstab
/etc/fstab
shitou@aishitou:~$ cp /etc/fstab /etc/fstab.bf
cp: cannot create regular file '/etc/fstab.bf': Permission denied
shitou@aishitou:~$ sudo cp /etc/fstab /etc/fstab.bf
shitou@aishitou:~$ sudo vim /etc/fstab
shitou@aishitou:~$ udo mount -a
Command 'udo' not found, but can be installed with:
sudo apt install udo
添加了 cpu 子系统的自动挂载配置（比如类似 none /sys/fs/cgroup/cpu cgroup defaults,cpu 0 0 的行），目的是让系统启动时自动挂载 cpu 子系统，避免每次手动挂载。
none /sys/fs/cgroup/cpu cgroup defaults,cpu,cpuacct 0 0
=========================================================================================================================
shitou@aishitou:~/monitor$ cat /etc/fstab
# /etc/fstab: static file system information.
#
# Use 'blkid' to print the universally unique identifier for a
# device; this may be used with UUID= as a more robust way to name devices
# that works even if disks are added and removed. See fstab(5).
#
# <file system> <mount point> <type> <options> <dump> <pass>
# / was on /dev/sda2 during curtin installation
/dev/disk/by-uuid/975f86b3-b8ea-4365-902d-5c61c0ad749c / ext4 defaults 0 1
# /boot was on /dev/sda3 during curtin installation
/dev/disk/by-uuid/069c7f52-682b-45f4-8c2e-2f2f38f9c60a /boot ext4 defaults 0 1
/dev/disk/by-uuid/8b865cd9-1d8f-48e5-bfe6-6a649909b853 none swap sw 0 0
# /boot/efi was on /dev/sda1 during curtin installation
/dev/disk/by-uuid/2D20-0B2E /boot/efi vfat defaults 0 1
none /sys/fs/cgroup/cpu cgroup defaults,cpu,cpuacct 0 0
-------------------------------------------------------------------------------------------
shitou@aishitou:~/monitor$ cat /etc/fstab.bf
# /etc/fstab: static file system information.
#
# Use 'blkid' to print the universally unique identifier for a
# device; this may be used with UUID= as a more robust way to name devices
# that works even if disks are added and removed. See fstab(5).
#
# <file system> <mount point> <type> <options> <dump> <pass>
# / was on /dev/sda2 during curtin installation
/dev/disk/by-uuid/975f86b3-b8ea-4365-902d-5c61c0ad749c / ext4 defaults 0 1
# /boot was on /dev/sda3 during curtin installation
/dev/disk/by-uuid/069c7f52-682b-45f4-8c2e-2f2f38f9c60a /boot ext4 defaults 0 1
/dev/disk/by-uuid/8b865cd9-1d8f-48e5-bfe6-6a649909b853 none swap sw 0 0
# /boot/efi was on /dev/sda1 during curtin installation
/dev/disk/by-uuid/2D20-0B2E /boot/efi vfat defaults 0 1
===========================================================================================================
-------------------------------------------------------------------------------
shitou@aishitou:~$ sudo mount -a
mount: /sys/fs/cgroup/cpu: none already mounted or mount point busy.
dmesg(1) may have more information after failed mount system call.
mount: (hint) your fstab has been modified, but systemd still uses
the old version; use 'systemctl daemon-reload' to reload.
shitou@aishitou:~$ mount | grep cgroup | grep cpu
cpuacct on /sys/fs/cgroup/cpuacct type cgroup (rw,relatime,cpuacct)
--------------------------------------------------------------------------------
执行 sudo mount -a 让 fstab 中的配置立即生效（无需重启），但报错 mount: /sys/fs/cgroup/cpu: none already mounted or mount point busy，说明：
要么 cpu 子系统已通过其他方式挂载（但实际 mount | grep cgroup | grep cpu 显示未挂载，更可能是冲突）；
要么与系统当前的 cgroup v2 机制冲突（cgroup v2 是统一挂载，不支持 v1 子系统的单独挂载）。
也就是从这个时候开始切换的
---------------------------------------------------------------------------------------------------
========================================================================================================
shitou@aishitou:~$ sudo nano /etc/fstab
shitou@aishitou:~$ sudo umount /sys/fs/cgroup/cpu
umount: /sys/fs/cgroup/cpu: not mounted.
尝试卸载/sys/fs/cgroup/cpu这个挂载点（取消该目录与 cgroupcpu子系统的关联）。
但提示not mounted，说明该目录当前并未被挂载，无需卸载。
shitou@aishitou:~$ sudo systemctl daemon-reload
重新加载配置文件
shitou@aishitou:~$ sudo umount /sys/fs/cgroup/cpu
umount: /sys/fs/cgroup/cpu: not mounted.
shitou@aishitou:~$ sudo systemctl daemon-reload
shitou@aishitou:~$ sudo mount -a
mount: /sys/fs/cgroup/cpu: none already mounted or mount point busy.
dmesg(1) may have more information after failed mount system call.
挂载/etc/fstab中所有未挂载的文件系统（按配置自动挂载）。
这里报错/sys/fs/cgroup/cpu: none already mounted or mount point busy，说明fstab中关于/sys/fs/cgroup/cpu的配置与当前系统（仍在使用 cgroup v2）冲突，无法挂载。
shitou@aishitou:~$ mount | grep cgroup | grep cpu
cpuacct on /sys/fs/cgroup/cpuacct type cgroup (rw,relatime,cpuacct)
查看系统中与cgroup和cpu相关的挂载记录。
输出cpuacct on /sys/fs/cgroup/cpuacct...说明cpuacct子系统已挂载，但cpu子系统仍未挂载（这正是cadvisor报错的核心原因）。
---------------------------------------------------------------------------------------------------
cpuacct：这是 cgroup（控制组）的一个子系统，主要用于CPU 资源使用的计量（accounting），即跟踪和统计进程或进程组消耗的 CPU 时间（用户态、内核态等）。
on /sys/fs/cgroup/cpuacct：表示 cpuacct 子系统被挂载到了 /sys/fs/cgroup/cpuacct 目录下。通过这个目录，用户可以配置和查看该子系统的相关参数（如进程组的 CPU 使用统计）。
type cgroup：说明这是一个 cgroup 类型的文件系统（cgroup 本质上是通过特殊文件系统实现的资源管理机制）。
(rw,relatime,cpuacct)：挂载选项。其中：
rw 表示该文件系统以可读写模式挂载；
relatime 表示文件访问时间会相对延迟更新（优化性能）；
cpuacct 是该 cgroup 子系统的标识。
---------------------------------------------------------------------------------------------
shitou@aishitou:~$ mount | grep cgroup2
cgroup2 on /sys/fs/cgroup type cgroup2 (rw,nosuid,nodev,noexec,relatime,nsdeleg ate,memory_recursiveprot)
检查系统是否在使用cgroup v2（新一代 cgroup 机制）。
输出cgroup2 on /sys/fs/cgroup...确认当前系统确实在使用cgroup v2，而cadvisor对其兼容性较差。
shitou@aishitou:~$ ls /sys/fs/cgroup/cgroup.controllers
/sys/fs/cgroup/cgroup.controllers
查看cgroup v2的控制器文件（cgroup.controllers是cgroup v2的标志性文件，记录了可用的资源控制子系统）
输出说明cgroup v2正常运行。
-------------------------------------------------------------------------------------
shitou@aishitou:~$ sudo nano /etc/default/grub
shitou@aishitou:~$ sudo cp /etc/default/grub /etc/default/grub.bf
shitou@aishitou:~$ sudo nano /etc/default/grubGRUB_CMDLINE_LINUX="systemd.unified_cgroup_hierarchy=0" 增加了着这个
目的是禁用cgroup v2
（cadvisor对v1兼容性更好）。 这里就又引出了一个问题，我的yaml文件是否是有问题，并且是否合理，
-----------------------------------------------------------------------------------
shitou@aishitou:~$ sudo update-grub
Sourcing file `/etc/default/grub'
Generating grub configuration file ...
Found linux image: /boot/vmlinuz-6.8.0-71-generic
Found initrd image: /boot/initrd.img-6.8.0-71-generic
Found linux image: /boot/vmlinuz-6.8.0-64-generic
Found initrd image: /boot/initrd.img-6.8.0-64-generic
Warning: os-prober will not be executed to detect other bootable partitions.
Systems on them will not be added to the GRUB boot configuration.
Check GRUB_DISABLE_OS_PROBER documentation entry.
Adding boot menu entry for UEFI Firmware Settings ...
done
根据/etc/default/grub的最新配置，重新生成GRUB启动菜单（使内核参数生效）。
输出显示已找到系统内核并更新配置，但新参数需重启系统后才会实际生效。