Description of problem: anaconda 30.25.4-1.fc30 crashes very early when testing the Fedora-30-20190408.n.0 nightly compose. from the console log ... Ý 648.256631¨ anacondaÝ2138¨: Starting installer, one moment... Ý 648.256680¨ anacondaÝ2138¨: anaconda 30.25.4-1.fc30 for Fedora 30 (pre-releas e) started. Ý 648.256719¨ anacondaÝ2138¨: * installation log files are stored in /tmp duri ng the installation Ý 648.256756¨ anacondaÝ2138¨: * if the graphical installation interface fails to start, try again with the Ý 648.256791¨ anacondaÝ2138¨: inst.text bootoption to start text installatio n Ý 648.256827¨ anacondaÝ2138¨: * when reporting a bug add logs from /tmp as sep arate text/plain attachments Ý 648.262648¨ anacondaÝ2138¨: Traceback (most recent call last): Ý 648.262694¨ anacondaÝ2138¨: File "/usr/sbin/anaconda", line 593, in <module > Ý 648.262729¨ anacondaÝ2138¨: initialize_network() Ý 648.262765¨ anacondaÝ2138¨: File "/usr/lib64/python3.7/site-packages/pyana MORE... DEVEL2 ================================================================================ onda/network.py", line 307, in initialize_network Ý 648.262799¨ anacondaÝ2138¨: Ýdev.device_name for dev in get_supported_dev ices()¨) Ý 648.262842¨ anacondaÝ2138¨: File "/usr/lib64/python3.7/site-packages/pyanac onda/network.py", line 557, in get_supported_devices Ý 648.262878¨ anacondaÝ2138¨: for device in network_proxy.GetSupportedDevic es()¨ Ý 648.262912¨ anacondaÝ2138¨: File "/usr/lib/python3.7/site-packages/pydbus/p roxy_method.py", line 102, in __call__ Ý 648.262949¨ anacondaÝ2138¨: raise error Ý 648.262983¨ anacondaÝ2138¨: pyanaconda.modules.common.errors.DBusError: Argum ent 0 does not allow None as a value Version-Release number of selected component (if applicable): anaconda 30.25.4-1.fc30 How reproducible: 100% Steps to Reproduce: 1. python ./nb.py --cmdline "" --prepend_cmdline "" --ks_edit "no" --ks_location "http" --ks_packageset "core" --ks_part "autopart" --ks_repo "" --lang "yes" --method "text" --type "rel-eng" --url "https://kojipkgs.fedoraproject.org/compose/branched/Fedora-30-20190408.n.0/compose/Everything/s390x/os/" --use_kickstart "yes" --reboot "yes" Additional information: This looks more as a generic issue than a s390x specific one.
Created attachment 1554245 [details] anaconda.log
Created attachment 1554246 [details] dbus.log
Created attachment 1554247 [details] program.log
Created attachment 1554248 [details] storage.log
Created attachment 1554249 [details] syslog
Proposed as a Freeze Exception for 30-final by Fedora user sharkcz using the blocker tracking app because: The crash makes Fedora 30 for s390x uninstallable and because we are alternative arch, proposing as FreezeException.
I think we are seeing the same issue as in https://bugzilla.redhat.com/show_bug.cgi?id=1695967#c4 here. It should be fixed by https://github.com/rhinstaller/anaconda/pull/1941. Or at least it should give us some more debugging info. If you want you can try the fix with this updates image (for the anaconda 30.25.4-1.fc30): https://rvykydal.fedorapeople.org/updates.nmdevicenone.img If you do so, could you please post /tmp/syslog (or /var/log/anaconda/syslog) ?
Created attachment 1554274 [details] syslog
With the updates.img the problem goes away. Attached is /tmp/syslog, because the installation failed due an infra issue (can't download a rpm). Giving it another try :-)
All looks good after another try with the updates.img applied.
Dan, thank you for checking. From the log it seems that as in https://bugzilla.redhat.com/show_bug.cgi?id=1695967#c13 the error condition didn't occur this time. But the PR has been merged so if it appears next time, we'll have at least better logging info.
Discussed during the 2019-04-15 blocker review meeting: [1] The decision to classify this bug as an "AcceptedFreezeException" was made as this causes installs to fail on s390 and clearly cannot be fixed with an update. [1] https://meetbot.fedoraproject.org/fedora-blocker-review/2019-04-15/f30-blocker-review.2019-04-15-16.03.txt
It really seems the proposed fix is correct and should resolve this, so I think we should close it. Dan, please let us know if you still see issues along these lines with F30 Final.
Rechecked with F-30 GA and it's been fixed indeed.