Since Fedora-Rawhide-20190824.n.0, all traditional installer install tests in openQA seem to be failing with this error in anaconda.log: 05:56:33,634 CRT exception: Traceback (most recent call last): File "/usr/lib64/python3.8/site-packages/gi/overrides/BlockDev.py", line 967, in wrapped ret = orig_obj(*args, **kwargs) File "/usr/lib64/python3.8/site-packages/gi/overrides/BlockDev.py", line 545, in lvm_lvcreate return _lvm_lvcreate(vg_name, lv_name, size, type, pv_list, extra) gi.repository.GLib.GError: g-io-error-quark: Failed to call the 'LvCreate' method on the '/com/redhat/lvmdbus1/Vg/0' object: GDBus.Error:org.freedesktop.DBus.Python.dbus.exceptions.DBusException: ('com.redhat.lvmdbus1.Vg', 'Exit code 3, stderr = Size requires number argument.\n Invalid argument for --size: dbus.UInt64(8585740288)B\n Error during parsing of command line.\n') (36) This is not happening in F31, so I'm gonna guess this is a Python 3.8 issue as that's the major difference between Rawhide and F31 ATM. Proposing as an F32 Beta blocker as this breaks all kinds of criteria obviously.
> This is not happening in F31, so I'm gonna guess this is a Python 3.8 issue as that's the major difference between Rawhide and F31 ATM. Yes, it's a Python 3.8 related issue. It's either bug in python-dbus or in lvmdbusd, see https://bugzilla.redhat.com/show_bug.cgi?id=1745597 Basically the `__str__` method of some python-dbus objects changed (or more precisely `__str__` implementation of Python object changed in 3.8) and `str(dbus.UInt64(8585740288)` no longer returns a number. I'm not closing this as a duplicate because I'm not sure if it's ok to close a Fedora blocker bug as a duplicate of a bug for another product, but this won't be fixed in libblockdev.
*** Bug 1745980 has been marked as a duplicate of this bug. ***
We'll need to keep a Fedora product bug open, yeah, because the closing of the bug must be linked to a Fedora update going stable. We can reassign this to the lvm2 package, though. I'll backport the patch.
https://koji.fedoraproject.org/koji/buildinfo?buildID=1365949 Should be fixed in next compose.
*** This bug has been marked as a duplicate of bug 1745597 ***
The reason why this should *not* be closed as a duplicate of that bug was explained in comment #3. It is necessary for the Fedora blocker tracking process.
Fix confirmed in Fedora-Rawhide-20190828.n.0 tests - https://openqa.fedoraproject.org/tests/overview?distri=fedora&version=Rawhide&build=Fedora-Rawhide-20190828.n.0&groupid=1 . This can be closed now.