Release 16.04, kernel 4.4.0-124-generic
There's an issue in kernel's NBD module which prevents some larger images to be correctly "connected" , largely described here: https://github.com/NetworkBlockDevice/nbd/issues/44
There is a small patch here that was accepted into mainline 4.10: https://www.spinics.net/lists/linux-block/msg07060.html
This is a regression from the previous LTS 3.13 kernel.
----
Here is a small example of the faulty behaviour:
# qemu-img create -f qcow2 test.img 1100G Formatting 'test.img', fmt=qcow2 size=1181116006400 cluster_size=65536 lazy_refcounts=off refcount_bits=16 # qemu-nbd -c /dev/nbd0 test.img # blockdev --getsize64 /dev/nbd0 18446743055802302464
The correct response would be 1181116006400; this breaks most tools and makes the image unusable, e.g.
# fdisk -l /dev/nbd0 fdisk: cannot open /dev/nbd0: Invalid argument
Release 16.04, kernel 4.4.0-124-generic
There's an issue in kernel's NBD module which prevents some larger images to be /github. com/NetworkBloc kDevice/ nbd/issues/ 44
correctly "connected" , largely described here:
https:/
There is a small patch here that was accepted into mainline 4.10: /www.spinics. net/lists/ linux-block/ msg07060. html
https:/
This is a regression from the previous LTS 3.13 kernel.
----
Here is a small example of the faulty behaviour:
# qemu-img create -f qcow2 test.img 1100G 02464
Formatting 'test.img', fmt=qcow2 size=1181116006400 cluster_size=65536 lazy_refcounts=off refcount_bits=16
# qemu-nbd -c /dev/nbd0 test.img
# blockdev --getsize64 /dev/nbd0
184467430558023
The correct response would be 1181116006400; this breaks most tools and makes the image unusable, e.g.
# fdisk -l /dev/nbd0
fdisk: cannot open /dev/nbd0: Invalid argument