Bug 1797986

Created attachment 1657529 [details]
Nmstate test log

We cannot reproduce it manually. I've attached the tests logs.

I'm not very familiar with Python and nmstate code, but from what I
understand the problem seems to be in the test.

test_interface_mtu_change_with_modify() calls _modify_interface()
which does this:

    conn = nm.connection.ConnectionProfile()
    conn.import_by_id(ETH1)
    settings = _create_iface_settings(wired_state, conn)
    new_conn = nm.connection.ConnectionProfile()
    new_conn.create(settings)
    conn.update(new_conn)
    conn.commit(save_to_disk=False)

    nmdev = nm.device.get_device_by_name(ETH1)
    apply_operation(nmdev, conn.profile)

so, it enqueues the commit/reapply operations in the main loop and
returns. Since the function is decorated with @nmclient_context, the
NMClient gets destroyed as soon as the function returns; therefore
when commit() later is dequeued and executed, it fails because the
RemoteConnection object doesn't have an associated client.

I haven't investigated how this worked with NM 1.20, but this seems to
me a bug in the nmstate test implementation. Fernando, what do you
think?

Changed to nmstate

(In reply to Beniamino Galvani from comment #2)
> I'm not very familiar with Python and nmstate code, but from what I
> understand the problem seems to be in the test.
> 
> test_interface_mtu_change_with_modify() calls _modify_interface()
> which does this:
> 
>     conn = nm.connection.ConnectionProfile()
>     conn.import_by_id(ETH1)
>     settings = _create_iface_settings(wired_state, conn)
>     new_conn = nm.connection.ConnectionProfile()
>     new_conn.create(settings)
>     conn.update(new_conn)
>     conn.commit(save_to_disk=False)
> 
>     nmdev = nm.device.get_device_by_name(ETH1)
>     apply_operation(nmdev, conn.profile)
> 
> so, it enqueues the commit/reapply operations in the main loop and
> returns. Since the function is decorated with @nmclient_context, the
> NMClient gets destroyed as soon as the function returns; therefore
> when commit() later is dequeued and executed, it fails because the
> RemoteConnection object doesn't have an associated client.
> 
> I haven't investigated how this worked with NM 1.20, but this seems to
> me a bug in the nmstate test implementation. Fernando, what do you
> think?

I think you are right. This test could be changed soon as nmstate-0.3 will contain big changes on the Mainloop topic. I will re-check it then.

Not a issue any more as nmstate has changed the use of MainContext.


The new test case name is `test_interface_mtu_change` in case QE want to try it out.

Verified with versions:
nmstate-0.3.3-2.el8.noarch
NetworkManager-1.26.0-1.el8.x86_64
DISTRO=RHEL-8.3.0-20200716.n.0
Linux hp-dl388g8-04.rhts.eng.pek2.redhat.com 4.18.0-226.el8.x86_64 #1 SMP Wed Jul 15 07:40:39 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux


[10:52:09@hp-dl388g8-04 ~/nmstate/tests/integration/nm]0# pytest wired_test.py -vv
============================================================================================================ test session starts =============================================================================================================
platform linux -- Python 3.6.8, pytest-5.4.3, py-1.9.0, pluggy-0.13.1 -- /usr/bin/python3.6
cachedir: .pytest_cache
RPMs: NetworkManager-1.26.0-1.el8.x86_64
OS: Red Hat Enterprise Linux 8.3 Beta (Ootpa)
nmstate: nmstate-0.3.3-2.el8.noarch

rootdir: /root/nmstate, inifile: tox.ini
collected 2 items                                                                                                                                                                                                                            

wired_test.py::test_interface_mtu_change PASSED                                                                                                                                                                                        [ 50%]
wired_test.py::test_interface_mac_change_with_modify PASSED                                                                                                                                                                            [100%]

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (nmstate bug fix and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2020:4696