When parsing "xdev"-kind devices, do not assume that the partition
number can be converted to integer: re_xdev accepts an empty part of the
partition number, so just handle as it is, as string.
This fixes a regression due to the conversion of the inspection code to
OCaml, as the old C version did not have this issue.
(cherry picked from commit cf6b527824)
Use a better icon for RHEL guests, still provided by redhat-logos (or
equivalent in downstream distributions), and which fits a better
definition of logo for the distribution.
Thanks to Ray Strode for the hints.
(cherry picked from commit c648052690)
We've been carrying this exact patch in RHEL 7 for several years. It
reverts the change made in 2014 where we switched to using the virbr0
bridge for libguestfs networking instead of SLIRP. We thought SLIRP
was going to become unsupported in qemu, but recently there have been
more encouraging signs since it looks like SLIRP will be spun off as a
separate project, running as a modular process and properly secured
and supported.
This reverts commit 224de20b9a.
(cherry picked from commit 492a945791)
So far RBufferOut return values, and FBuffer struct fields are 'str' on
all the versions of Python. Python 3 distinguishes between 'str'
(unicode strings), and 'bytes', with 'str' no more able to hold
arbitrary data.
For this reason, switch the return value of RBufferOut functions, and
FBuffer struct fields to bytes on Python 3: while this is a potentially
incompatibile change, this is the only way to handle safely sequences
of arbitrary bytes.
(cherry picked from commit 0ee02e0117)
Make sure to reference the arguments, to make sure they are kept alive
during the function call; this is visible when setting an event handler
for the CLOSE event, and testing it with Python 3.
This does not seem to create a memory leak e.g. with Python 2.
Also, switch away from the quasi-internal PyEval_CallObject to the
public PyObject_CallObject, which takes care of doing safety checks.
(cherry picked from commit 85235aec83)
Remove "Allow virtual machine download", added with
commit 37955f14aa, because it does not
seem to be actually needed.
Related: RHBZ#1530967
(cherry picked from commit 5e5896752e)
Do not assume that the Python plugin of nbdkit has the same name of the
Python interpreter.
Use the default upstream name of nbdkit to identify it; downstream
distributions must adjust this variable, in case they rename the Python
plugin of nbdkit.
(cherry picked from commit 6b80c5fb51)
Add few more characters for the devices of resume= entries in the
command line of grub: this way it is possible to match also /dev/mapper
devices.
This should require no further processing, since the names of the
/dev/mapper devices do not change after the conversion.
(cherry picked from commit 50b08fd84b)
The previous support documentation said that for UEFI guests, when
using RHEL ≥ 7.3 host, you must use qemu-kvm-rhev. For RHEL 8 this is
inaccurate since qemu-kvm-rhev no longer exists.
To fix this I've dropped the whole sentence. It only applies to the
downstream product (RHEL) and so RHEL can add the right documentation
if it needs to.
Thanks: Ming Xie, Pino Toscano.
(cherry picked from commit 0e6bbcee7e)
See:
https://bugzilla.redhat.com/show_bug.cgi?id=1584678#c15
Fixes commit bcdbe6405c. However this
bug was copied directly from old virt-v2v which did the same thing
(from lib/Sys/VirtConvert/Converter/Windows.pm):
echo installing rhev-apt >>log.txt
"rhev-apt.exe" /S /v /qn >>log.txt
Thanks: Lev Veyde
(cherry picked from commit e12c56176a)
QEMU for x86 supports two machine types, "pc" (emulating the ancient
Intel i440FX chipset originally used by the Pentium Pro), and "q35"
(https://wiki.qemu.org/Features/Q35).
Currently virt-v2v does not set any machine type, so libvirt or the
target hypervisor will choose some default, probably i440fx. The
latest advice from the QEMU and libvirt communities is not to rely on
the default machine type but to specify what we need explicitly, but
it may also be that the libvirt configuration file has been changed to
set the default machine type to Q35 (either by the distro or the end
user).
None of this matters for reasonably new guests since they can boot
with either chipset. However there are some very old guests (notably
Windows XP) which cannot handle Q35.
This commit changes virt-v2v so it always tries to specify the machine
type explicitly (assuming the target supports that, and not all of
them do). For x86_64 guests this patch always selects i440fx (pc).
In future we hope to get the correct machine type for the guest from
libosinfo but this is not available yet.
For non-x86 architectures we select the "virt" model which will
probably only work for AArch64. More work is needed for POWER.
For secure boot we still have to force the machine type to Q35. In a
future version this forcing can be removed since any guest which is
using secure boot will be new enough that it'll be using Q35 anyway
(on x86).
(cherry picked from commit 55879b2f2c)
Pass the LDFLAGS properly as arguments for the C linker when using
ocamlmklibs via the -ldopt option.
Followup of commit 34c23403c5.
(cherry picked from commit 4ff573c352)
After this commit, all annocheck errors are fixed except for:
Hardened: virt-get-kernel: MAYB: Gaps were detected in the annobin coverage. Run with -v to list.
After discussion with the annocheck maintainers this gap in coverage
(which corresponds to the OCaml runtime) seems to be caused either by
the runtime not being linked with the right flags, or might be a bug
in annocheck itself. In any case it's not something that can be
resolved within the scope of libguestfs.
(cherry picked from commit 34c23403c5)
OCaml has a small runtime which is statically linked into the virt
tools (providing things like GC and primitives). Since OCaml 4.03 it
has been possible to select variants of this runtime, one of which is
compiled with -fPIC, using ‘ocamlopt -runtime-variant _pic’.
This has performance implications on i686, but is relatively free on
other architectures. Since it (in theory) adds to the security of the
final binary this commit enables it whenever it is available.
(cherry picked from commit 09abb9c990)
The output method used the s_disk_id field assuming it was a unique,
monotonically increasing number counting from 0. However this is not
the case, the input method simply has to set s_disk_id to be unique
for each disk.
Fixes commit cc04573927.
Thanks: Xiaodai Wang
(cherry picked from commit 90c9c12258)
This device naming is mostly written by virt-p2v, so get the slot from
it directly without using the drive_index "decoding" function.
(cherry picked from commit c9abaa3e96)
Search for DC by 'storage.name=' to make it explicit.
I.e. "storage=" uses regex, so similar names can be
found in the search query. For example, searching for
a domain named FCSD, will find FCSD1 as well.
Due to a conflict with the IDs of the OVF standard, and the existing
implementation in ovirt-engine, the ID of QXL devices changed to a
different value.
As a consequence, change the ResourceType of QXL devices, but only in
OVirt flavour to avoid breaking vdsm mode.
See: https://bugzilla.redhat.com/show_bug.cgi?id=1598715#c5
(cherry picked from commit aa9e18f6d1)
For direct upload, a suitable host must be in status 'Up'
and belong to the same datacenter as the created disk.
Added these criteria to the host search query.
(cherry picked from commit 4ed1bc5a79)
The optimization to start the transfer on the local host makes sense
only when using the rhv-direct=true option. When using a proxy, let the
engine choose a host.
(cherry picked from commit 891b5a0ec0)
When optimizing the connection using unix socket, we handle these cases:
- The local host is not an oVirt host (no /etc/vdsm/vdsm.id).
- The local host is an oVirt host, but is not registered with engine.
- Creating UnixHTTPConnection() fails. Unlikely and probably a bug in
the plugin, but we can recover by using the https connection.
The current code handle these cases silently, making it harder to
understand why the unix socket optimization did no happen. Add debug
message to make this clear.
Also comment in the error handler why we take this path instead of
failing the operation.
(cherry picked from commit f5442d2f04)
Old imageio proxy was using Authorization header for GET and PUT
requests. Remove unneeded authorization when sending OPTIONS request.
Remove unneeded duplicated comments about authorization for old
imageio, and replace them with a comment when we set needs_auth.
(cherry picked from commit d5f36bacf9)
In the case where virt-v2v runs on the same server as the imageio
daemon that we are talking to, it may be possible to optimize access
using a Unix domain socket.
This is only an optimization. If it fails or if we're not running on
the same server it will fall back to the usual HTTPS over TCP
connection.
Thanks: Nir Soffer, Daniel Erez.
(cherry picked from commit b7a2e6270d)
Previously we lazily requested the server options in the can_*
callbacks. The can_* callbacks are always called by nbdkit straight
after open, so this just adds complexity for no benefit. This change
simply makes the code always fetch the server options during the open
callback.
This is — functionally at least — mostly just refactoring. However I
also added a useful debug message so we can see what features the
imageio server is offering.
(cherry picked from commit a1e1f6ec88)
"http" and "transfer" variables were missing in emulate_zero, so the
code would fail with NameError. This can happen only when communicating
with old imageio versions not supporting the "zero" feature.
Testing with qemu-img 2.12 show that we never send emulated zero request
because of the highestwrite mechanism, but it can break with older
qemu-img-rhev used on RHEL.
(cherry picked from commit 0ae61ce99c)
Python manual warns[1]:
Note that you must have read the whole response before you can send
a new request to the server.
The reason for this warning is exposed only when the server is using
keep alive connections. When the response is not read, sending a new
request will fail with:
httplib.ResponseNotReady
Even if Content-Length was 0 or the request has no content. The failure
looks like this when using --verbose:
nbdkit: python[1]: debug: zero count=33554432 offset=0 may_trim=1 fua=0
nbdkit: python[1]: debug: zero count=33554432 offset=33554432 may_trim=1 fua=0
nbdkit: python[1]: error: /home/nsoffer/src/libguestfs/tmp/rhvupload.Au2B5I/rhv-upload-plugin.py: zero: error: Request-sent
nbdkit: python[1]: debug: sending error reply: Input/output error
qemu-img: error writing zeroes at offset 0: Input/output error
Change all requests to read the whole response.
Tested with imageio patch supporting keep alive connections:
https://gerrit.ovirt.org/#/c/92296/
[1] https://docs.python.org/3.8/library/http.client.html#http.client.HTTPConnection.getresponse
(cherry picked from commit f4e0a8342d)
When sending request with small or no payload, it is simpler and
possibly more efficient to use the high level HTTPSConnection.request(),
instead of the lower level APIs.
The only reason to use the lower level APIs is to avoid copying the
payload, or on python 2, to use a bigger buffer size when streaming a
file-like object.
(cherry picked from commit 77a412c0a1)
Output also the osinfo guess for each OS in the generated XML output;
adapt the RNG schema, and the test data to it.
(cherry picked from commit 2f8ec91fc1)
This increases the inactivity timeout for transfers from the default
(60 seconds in oVirt < 4.2.3, 600 seconds in >= 4.2.3), up to 1 hour,
so that we should never hit it for ordinary transfers.
Note this requires oVirt >= 4.2.3. The corresponding oVirt fix was
https://bugzilla.redhat.com/1563278
I also replaced the deprecated ‘image’ parameter with ‘disk’.
Thanks: Nir Soffer, Daniel Erez.
(cherry picked from commit f25404c65f)
Add an helper to convert an OVF flavour to string, and use it in
-o vdsm to print the actual value of the vdsm-ovf-flavour option.
Thanks to Ming Xie.
(cherry picked from commit c0da02f90d)
Use the $(srcdir) variable where needed, to make sure it builds also
with srcdir != builddir.
Furthermore, make sure that the OCaml dependencies calculation depend on
the generated output_rhv_upload_*_source.ml files, otherwise there will
be incomplete OCaml rules for them in the generated .depend.
Fixes commit cc04573927.
(cherry picked from commit a4e3b7c059)
Because we checked the limitation of raw+sparse during command line
processing, it had the effect of forcing you to use ‘-of raw’ even if
the input was already in raw format. Move the checking to the output
mode to avoid this.
Fixes commit cc04573927.
(cherry picked from commit b3769afaec)
This adds a new output mode to virt-v2v. virt-v2v -o rhv-upload
streams images directly to an oVirt or RHV >= 4 Data Domain using the
oVirt SDK v4. It is more efficient than -o rhv because it does not
need to go via the Export Storage Domain, and is possible for humans
to use unlike -o vdsm.
The implementation uses the Python SDK (‘ovirtsdk4’ module). An
nbdkit Python 3 plugin translates NBD calls from qemu into HTTPS
requests to oVirt via the SDK.
(cherry picked from commit cc04573927)
