similar to: [PATCH 0/2] v2v: rhv-upload-plugin: Improve error handling

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. --- v2v/rhv-upload-plugin.py | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) diff --git a/v2v/rhv-upload-plugin.py

Replace python 2 only "buffer" with "memoryview". Falling back to emulated zero would fail with: NameError: name 'buffer' is not defined I did not test the changed code but it was not tested before so it is unlikely to be worse. Detected by pylint. --- v2v/rhv-upload-plugin.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git

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. --- v2v/rhv-upload-plugin.py | 35 ++++++++++++++++-------------------

Only compile tested. I'm keeping this as a separate patch for the moment until the relevant change on the oVirt side goes upstream. Rich.

v8 was here: https://www.redhat.com/archives/libguestfs/2018-April/msg00022.html v8 -> v9: - Addresses the only feedback from Tomáš. Rich.

v7 was here: https://www.redhat.com/archives/libguestfs/2018-March/msg00143.html Since then: - Earlier patches are now upstream. - The to-do list is moved from the commit message to the TODO file. - This version forces -of raw + -oa sparse and gives an error in any other mode. We intend to lift these restrictions later. - Tested against latest imageio which supports longer timeouts,

Fixes the majority of the pycodestyle issues in the Python scripts, and fix the existing test-v2v-python-syntax.sh to use pycodestyle to actually perform style checks. Pino Toscano (6): PEP 8: adapt whitespaces in lines PEP 8: move imports at the top PEP 8: adapt empty lines tests: find all the Python scripts for syntax checks -o rhv-upload: remove unused Python imports Revamp check

With python 3, we have a nicer way to handle socket.error with errno set to EPIPE (or ESHUTDOWN). This is also more correct since in some cases (that I could not reproduce yet with v2v), using e[0] with BrokenPipeError will fail with: >>> OSError(errno.EPIPE, "Broken pipe")[0] Traceback (most recent call last): File "<stdin>", line 1, in <module>

v1 was here: https://www.redhat.com/archives/libguestfs/2018-June/msg00099.html v2: - Just fixes the two problems noted in the review of the previous version. Rich.

On Thu, Mar 22, 2018 at 5:25 PM Richard W.M. Jones <rjones@redhat.com> wrote: > PROBLEMS: > - -of qcow2 does not work, with multiple problems > * needs to set NBD size to something larger than virtual size > How is this related to the actual file size specified when you create a disk? In block storage, qcow2 image is always on a thin provisioned disk, which in oVirt is a

v5 was here: https://www.redhat.com/archives/libguestfs/2018-March/msg00032.html There is only a single patch in this version because the other patches went upstream. This patch adds the virt-v2v -o rhv-upload mode (https://bugzilla.redhat.com/show_bug.cgi?id=1557273). Compared to v5, this adds the ability to make zero, trim and flush requests to the oVirt imageio server

http was 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. --- v2v/rhv-upload-plugin.py | 2 ++ 1 file

These two patches add support for using a Unix domain socket to directly access imageio in the case where imageio is running on the conversion host (usually that means virt-v2v is running on the RHV node and something else -- eg. CFME scripts -- arranges that the RHV node is the same one running imageio). Conversions in the normal case are not affected - they happen over TCP as usual. This was

v6 was here: https://www.redhat.com/archives/libguestfs/2018-March/msg00126.html This makes a number of significant changes: - Input and output options now use a uniform set of -io and -oo parameters. - For -o rhv-upload, we use ‘-oo rhv-cafile=/tmp/ca.pem’ etc. The ‘--rhv*’ options have been dropped. - Rearranges the documentation. - As before includes (untested) support for zero,

Nir, can you confirm this is correct for RHEL 7 / Python 2? Rich.

v1 was here: https://www.redhat.com/archives/libguestfs/2018-June/msg00015.html v2 fixes all the issues raised in Nir's review. I tested it against ovirt-engine-4.2.4.1-1.el7.noarch and it works, although I wasn't easily able to trigger the error path so that code is not really tested (except for syntax checking). Rich.

Remove some instances of "transfer" and "transfer_service" that are not used in current code. --- v2v/rhv-upload-plugin.py | 4 ---- 1 file changed, 4 deletions(-) diff --git a/v2v/rhv-upload-plugin.py b/v2v/rhv-upload-plugin.py index 10887c031..5c036c46a 100644 --- a/v2v/rhv-upload-plugin.py +++ b/v2v/rhv-upload-plugin.py @@ -263,7 +263,6 @@ def request_failed(h, r, msg):

The oVirt server may fail a PUT request before reading all request body. However before closing the connection, it writes a detailed error response, that will make debugging issues like expired tickets much easier to handle. If we don't get the response and log the error, the error may be lost when the daemon log is rotated. Change pwrite() and emulate_zero() to get the response after EPIPE,

Mainly minor fixes and code cleanups over the v4 patch. There are still several problems with this patch, but it is in a reviewable state, especially the Python code. Rich.

v2 was here: https://www.redhat.com/archives/libguestfs/2018-June/msg00109.html v3: - Added/fixed all suggestions from Nir in previous review. Q: I wasn't sure if we still need the "UnsupportedError" class so I left it in. Q: Does the Unix socket always have the same name? What happens if there's more than one transfer happening? I tested this both ways, and it worked both