On Thursday, 16 December 2021 14:11:34 CET Nir Soffer wrote:
On Thu, Dec 16, 2021 at 1:36 PM Milan Zamazal
<mzamazal(a)redhat.com> wrote:
> Hi,
>
> it seems a recent commit causes a storage test failure on GitHub:
> ================================== FAILURES
> =================================== ________________
> test_prepare_connection_without_initiator_name ________________>
> def test_prepare_connection_without_initiator_name():
> con_def = [{
>
> "password": "password",
> "port": "3260",
> "iqn": "iqn.2016-01.com.ovirt:444",
> "connection": "192.168.1.2",
> "ipv6_enabled": "false",
> "id": "994a711a-60f3-411a-aca2-0b60f01e8b8c",
> "user": "",
> "tpgt": "1",
>
> }]
>
> > conn = storageServer._prepare_connections(sd.ISCSI_DOMAIN,
> > con_def)
>
> E AttributeError: module 'vdsm.storage.storageServer' has no
> attribute '_prepare_connections'>
> Could you please fix it quickly or disable the test?
>
> @all: Beware that CI on Vdsm no longer runs the tests and patches must
> be tested on GitHub before merging.
Thanks for reporting, fixed in
https://gerrit.ovirt.org/c/vdsm/+/118071
Fix is now merged in the master. Sorry for that and thanks Nir for the fix.
I missed this error in the morning as there is another failure (CodeQL fails)
and the tests didn't run after the merge.
Vojta
I guess this is a side effect of testing only the top patch in
github.
It is much faster but misses inconsistencies in the series.
Once we move to github this should not happen since in github
you cannot merge a partial PR. It can still happen if we checkout
random patch from a PR that was never tested alone, but this is
not very interesting.
Nir