docker服务无法启动的解决方案(layer does not exist)

近期一台服务器上的docker无法启动,查看日志发现报告的是:msg=”Error starting daemon: layer does not exist”

现象

反复重启也无法启动,日志里面可以看到:

msg=”Error starting daemon: layer does not exist”

其实有的时候可能还会有其他的一些文件系统错误啥的日志。

解决

到网上搜索后发现似乎没啥好的办法,只能是清空/var/lib/docker目录了,实际上官方还提供了一个略微安全的删除脚本:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
#!/bin/sh
set -e

dir="$1"

if [ -z "$dir" ]; then
{
echo 'This script is for destroying old /var/lib/docker directories more safely than'
echo ' "rm -rf", which can cause data loss or other serious issues.'
echo
echo "usage: $0 directory"
echo " ie: $0 /var/lib/docker"
} >&2
exit 1
fi

if [ "$(id -u)" != 0 ]; then
echo >&2 "error: $0 must be run as root"
exit 1
fi

if [ ! -d "$dir" ]; then
echo >&2 "error: $dir is not a directory"
exit 1
fi

dir="$(readlink -f "$dir")"

echo
echo "Nuking $dir ..."
echo ' (if this is wrong, press Ctrl+C NOW!)'
echo

( set -x; sleep 10 )
echo

dir_in_dir() {
inner="$1"
outer="$2"
[ "${inner#$outer}" != "$inner" ]
}

# let's start by unmounting any submounts in $dir
# (like -v /home:... for example - DON'T DELETE MY HOME DIRECTORY BRU!)
for mount in $(awk '{ print $5 }' /proc/self/mountinfo); do
mount="$(readlink -f "$mount" || true)"
if dir_in_dir "$mount" "$dir"; then
( set -x; umount -f "$mount" )
fi
done

# now, let's go destroy individual btrfs subvolumes, if any exist
if command -v btrfs > /dev/null 2>&1; then
root="$(df "$dir" | awk 'NR>1 { print $NF }')"
root="${root#/}" # if root is "/", we want it to become ""
for subvol in $(btrfs subvolume list -o "$root/" 2>/dev/null | awk -F' path ' '{ print $2 }' | sort -r); do
subvolDir="$root/$subvol"
if dir_in_dir "$subvolDir" "$dir"; then
( set -x; btrfs subvolume delete "$subvolDir" )
fi
done
fi

# finally, DESTROY ALL THINGS
( set -x; rm -rf "$dir" )

将该脚本保存到本地后运行,清空/var/lib/docker目录,重启docker服务,果然好使了。

参考

坚持原创技术分享,您的支持将鼓励我继续创作!