similar to: [PATCH] v2v: Allow output modes to rewrite disk copying

This changes the infrastructure to allow the target_file to be a QEMU URI. Rich.

We previously implicitly supported writing to block devices instead of local files, but there were several problems: * Block devices could be deleted, especially if virt-v2v failed during a conversion. * Block devices could be overwritten by a file with the same name, although I believe this is just an observed consequence of the previous point, or at least I was not able to reproduce this

Factor out copying the overlays to final disk images into a separate function. Signed-off-by: Roman Kagan <rkagan@virtuozzo.com> --- v2v/v2v.ml | 227 +++++++++++++++++++++++++++++++------------------------------ 1 file changed, 114 insertions(+), 113 deletions(-) diff --git a/v2v/v2v.ml b/v2v/v2v.ml index afffde2..703038c 100644 --- a/v2v/v2v.ml +++ b/v2v/v2v.ml @@ -110,121 +110,9 @@ let

On Tue, Sep 01, 2020 at 08:55:38PM +0300, Nir Soffer wrote: ... > > Note this commit intentionally does not prevent you from writing qcow2 > > to a block device. RHV uses this so it is a thing that people do. ... > > +Note that you must create block devices of the correct size, and you > > +need to use I<-of raw> since other output formats would not normally >

On Tue, Sep 1, 2020 at 4:56 PM Richard W.M. Jones <rjones@redhat.com> wrote: > > We previously implicitly supported writing to block devices instead of > local files, but there were several problems: > > * Block devices could be deleted, especially if virt-v2v failed during > a conversion. > > * Block devices could be overwritten by a file with the same name, >

This patch series fixes various issues in the rhv-upload output mode: - properly find and use RHV resources - cleanup orphan disks, and possibly the current disk transfer on failure In reality, the first 4 patches deal with resources, and the other 4 with cleanups. The latter block can be theoretically sent alone -- I just happened to start working on it as part of my "let's fix

It will be used to do extra checks in the output before copying the disks. Partially revert commit 3bafec4e693a25ef1c84abc0fd1bc3251862c7de. --- v2v/output_glance.ml | 2 +- v2v/output_json.ml | 2 +- v2v/output_libvirt.ml | 2 +- v2v/output_local.ml | 2 +- v2v/output_null.ml | 2 +- v2v/output_openstack.ml | 2 +- v2v/output_qemu.ml | 2 +- v2v/output_rhv.ml

Add a simple method in the Output class to do work right after a disk was successfully copied. --- v2v/types.ml | 1 + v2v/types.mli | 4 ++++ v2v/v2v.ml | 9 ++++++++- 3 files changed, 13 insertions(+), 1 deletion(-) diff --git a/v2v/types.ml b/v2v/types.ml index 77f879200..714b30014 100644 --- a/v2v/types.ml +++ b/v2v/types.ml @@ -521,6 +521,7 @@ class virtual output = object method

Add support for qcow2 disk format, enabled by imageio NBD backend in RHV 4.3. To use this feature manually, you can run virt-v2v with "-of qcow2". Here is example run: Source disk: $ qemu-img info /var/tmp/fedora-30.img image: /var/tmp/fedora-30.img file format: raw virtual size: 6 GiB (6442450944 bytes) disk size: 1.15 GiB virt-v2v: $ ./run virt-v2v \ -v \ -i disk

This series is a second attempt to add a mode of virt-v2v operation where it leaves the config and disk image conversion, rollback on errors, registering with the destination hypervisor, etc. to a third-party toolset, and performs only tuning of the guest OS to run in the KVM-based hypervisor. The first 14 patches are just refactoring and rearrangement of the code, factoring the implementation

This patch series adds stable bus and slot numbers for removable drives (CDs and floppies) when the guest is converted using virt-v2v or virt-p2v. Previously we were a bit random about this. After this patch series, the bus and slot numbers and preserved if at all possible. BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1238053 Rich.

This adds --print-target. In addition, v2 provides a machine-readable version (in JSON format). All of the record -> JSON boilerplate in this patch could be eliminated if we moved the baseline to OCaml 4.02. Rich.

Commit 4699c7b6e126e07c95b67fb95df58aed87a680dd converted the null output to use the null-co qemu driver with a JSON URL syntax -- especially the latter is only available in newer versions of qemu. Even if this output mode is mostly for testing, check at runtime whether the null-co + JSON way is possible, falling back to the creation of thrown-away temporary files as before. ---

This is analogous to --print-source, except that it prints the overlay and target disk information. The output looks like below. Note there is one overlay and one target section per disk. Overlay and Target information (--print-target option): overlay file: /home/rjones/d/libguestfs/tmp/v2vovlc687fe.qcow2 overlay device name: sda overlay virtual disk size: 6442450944

On Mon, Nov 25, 2019, 11:15 Richard W.M. Jones <rjones@redhat.com> wrote: > On Wed, Nov 20, 2019 at 03:06:55AM +0200, Nir Soffer wrote: > > diff --git a/v2v/v2v.ml b/v2v/v2v.ml > > index 03590c9e..58bb06c3 100644 > > --- a/v2v/v2v.ml > > +++ b/v2v/v2v.ml > > @@ -739,7 +739,9 @@ and copy_targets cmdline targets input output = > > | TargetURI

When using oVirt >= 4.3, we can enable the NBD based backend in imageio by specifying that we transfer raw data when creating a transfer. With the NBD backend, we can specify qcow2 disk format. I think we need a way to expose the capabilities of the output to the user. For example, can we use qcow2 format. Issues: - I hacked qemu-img convert command line to always use -O raw instead of

Besides, it doesn't need guestfs handle open so move this step earlier in the process. Signed-off-by: Roman Kagan <rkagan@virtuozzo.com> --- v2v/v2v.ml | 77 ++++++++++++++++++++++++++++++++------------------------------ 1 file changed, 40 insertions(+), 37 deletions(-) diff --git a/v2v/v2v.ml b/v2v/v2v.ml index 155eb83..4257b8d 100644 --- a/v2v/v2v.ml +++ b/v2v/v2v.ml @@ -58,6 +58,7

RHV was connecting to the VMware source in order to find out the virtual disk size of the source disk(s), duplicating logic that virt-v2v already provides. This makes that information available using a new ‘virt-v2v --print-target option’. Note in order to get all the information, this has to actually perform the conversion step, so it takes 30 seconds or so before we reach the point where the

On Tue, Sep 01, 2020 at 07:37:42PM +0100, Richard W.M. Jones wrote: > On Tue, Sep 01, 2020 at 08:55:38PM +0300, Nir Soffer wrote: > ... > > > Note this commit intentionally does not prevent you from writing qcow2 > > > to a block device. RHV uses this so it is a thing that people do. > ... > > > +Note that you must create block devices of the correct size, and

I'm interested to hear opinions on whether this makes the code clearer, or not. This is virt-v2v, but many other virt-* tools work the same way, and analogous changes could be made. Currently when command line argument parsing is done in 'cmdline.ml' the list of parsed parameters is passed to the main program in a very long tuple. Each parameter is strongly typed, but not named (so