qemu-img操作文件出现“Could not read snapshots: File too large”问题解决办法

因为kvm虚拟机磁盘满，准备扩容磁盘，但用qemu-img命令执行扩容扩容的时候出现了下面的问题

# qemu-img resize /data/AnolisOS.qcow2 +20G
qemu-img: Could not open '/data/AnolisOS.qcow2': Could not read snapshots: File too large

然后发现虚拟机也启动不了了

# virsh start AnolisOS
错误：开始域 AnolisOS 失败
错误：内部错误：qemu unexpectedly closed the monitor: 2022-09-05T08:57:21.971604Z qemu-kvm: -drive file=/data/AnolisOS.qcow2,format=qcow2,if=none,id=drive-virtio-disk0: could not open disk image /ssd/vm/host-38.AnolisOS.qcow2: Could not read snapshots: File too large

通过百度找到仅有的一篇，无头无尾 qemu-img报错 - 知了社区

果断挂梯子，Google之，资料多了些，好像是一个bug，有人已经遇到了 https://bugs.launchpad.net/qemu/+bug/1354167

2个解决方案，都要下载qemu源码，todd大神建议1.7.2版的

# wget https://download.qemu.org/qemu-1.7.2.tar.bz2
。。。。
# tar jxvf qemu-1.7.2.tar.bz2

1、参考百度来的方案使用test工具（bug回复中也有人回）

# ./tests/qemu-iotests/qcow2.py /data/AnolisOS.qcow2 dump-header
magic                     0x514649fb
version                   3
backing_file_offset       0x0
backing_file_size         0x0
cluster_bits              16
size                      21475270656
crypt_method              0
l1_size                   62
l1_table_offset           0x11a7c0000
refcount_table_offset     0x10000
refcount_table_clusters   1
nb_snapshots              2
snapshot_offset           0xb0000
incompatible_features     0x0
compatible_features       0x0
autoclear_features        0x0
refcount_order            4
header_length             104
 
# ./tests/qemu-iotests/qcow2.py /data/AnolisOS.qcow2 set-header nb_snapshots 0
# ./tests/qemu-iotests/qcow2.py /data/AnolisOS.qcow2 set-header snapshot_offset 0x0
# ./tests/qemu-iotests/qcow2.py /data/AnolisOS.qcow2 dump-header
magic                     0x514649fb
version                   3
backing_file_offset       0x0
backing_file_size         0x0
cluster_bits              16
size                      21475270656
crypt_method              0
l1_size                   62
l1_table_offset           0x11a7c0000
refcount_table_offset     0x10000
refcount_table_clusters   1
nb_snapshots              0
snapshot_offset           0x0
incompatible_features     0x0
compatible_features       0x0
autoclear_features        0x0
refcount_order            4
header_length             104

成功！

2、编译后使用编译后的执行文件转换（不要安装）

# ./configure && make

编译后执行

# ./qemu-img convert /data/AnolisOS.qcow2 /data/AnolisOS.fixed.qcow2
# mv /data/AnolisOS.qcow2 /data/AnolisOS.qcow2.bak && mv /data/AnolisOS.fixed.qcow2 /data/AnolisOS.qcow2

检查文件信息正常了，损害的时候info查看不了

# qemu-img info /data/AnolisOS.qcow2 
image: /data/AnolisOS.qcow2
file format: qcow2
virtual size: 20G (21475270656 bytes)
disk size: 25G
cluster_size: 65536
Format specific information:
    compat: 1.1
    lazy refcounts: false

 之后扩容就正常了

# qemu-img resize /data/AnolisOS.qcow2 +20G
Image resized.