[Bug 1000732] New: zypper dup from 13.2 to leap42.1 changes the ether names from enpXsX back to ethX but does not update network config
http://bugzilla.suse.com/show_bug.cgi?id=1000732 Bug ID: 1000732 Summary: zypper dup from 13.2 to leap42.1 changes the ether names from enpXsX back to ethX but does not update network config Classification: openSUSE Product: openSUSE Distribution Version: Leap 42.1 Hardware: Other OS: Other Status: NEW Severity: Major Priority: P5 - None Component: Network Assignee: bnc-team-screening@forge.provo.novell.com Reporter: rens.groenewegen@xs4all.nl QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- Created attachment 693890 --> http://bugzilla.suse.com/attachment.cgi?id=693890&action=edit all logs from /var/log/YaST2 zypper dup from 13.2 to leap42.1 changes the ether names from enpXsX back to ethX but does not update network config. system is a KVM host, 32GB, 8core AMD, one rtl and two intel gbit cards. I updated from 13.1 to 13.2 without issue. rebooted, all was well I then updated from 13.2 to leap42.1 all enpXs0 ETHER devices ( i have three network cards in this machine) were renamed to old-style ethX but the ifcfg-* files were not updated, resulting in non-reachability, no network fortunately I had a serial terminal, installed because of previous problems with networking (when configuring a bridge, it looses it already configured devices, this is one of the bugs we have in suse since 13.1, very annoying. ) this costs a lot of time, friends :-) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1000732 Chenzi Cao <chcao@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Assignee|bnc-team-screening@forge.pr |yast2-maintainers@suse.de |ovo.novell.com | -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1000732 http://bugzilla.suse.com/show_bug.cgi?id=1000732#c1 Michal Filka <mfilka@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mfilka@suse.com, | |wicked-maintainers@suse.de Flags| |needinfo?(wicked-maintainer | |s@suse.de) --- Comment #1 from Michal Filka <mfilka@suse.com> --- Well, you did "zypper dup", so yast is not involved in upgrade process. The issue most probably happens bcs Leap is more tightly coupled with SLE. In SLE we use old device naming scheme. @wicked team: Could you check it? Do you have some comments? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1000732 http://bugzilla.suse.com/show_bug.cgi?id=1000732#c2 rens groenewegen <rens.groenewegen@xs4all.nl> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(wicked-maintainer | |s@suse.de) | --- Comment #2 from rens groenewegen <rens.groenewegen@xs4all.nl> --- zypper dup is not yast. i know as much. however, yast is not capable of fixing the problem by itself. i had to manually adjust the stuff in /etc/syconfig/network in order to get it back on its feet again. perhaps you can find something in the logs why it could not do that. cheers, Rens -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1000732 http://bugzilla.suse.com/show_bug.cgi?id=1000732#c3 --- Comment #3 from rens groenewegen <rens.groenewegen@xs4all.nl> --- I just realized it looks like we have two issues leap42 is the follow up on 13.x, but uses pre-13.x handling of ethX naming conventino - yast2-lan is still making a mess of network config when having to rename interfaces or finds interfaces with other names than expected. or actually, it does not even notice. would be practical to use the MAC addres to record / to make sure what is what for at least the physical interfaces ?? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1000732 http://bugzilla.suse.com/show_bug.cgi?id=1000732#c4 Steffen Winterfeldt <snwint@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |lnussel@suse.com, | |snwint@suse.com Flags| |needinfo?(lnussel@suse.com) --- Comment #4 from Steffen Winterfeldt <snwint@suse.com> --- Yes, we've put us into a pretty mess here ourselves. 13.2: enp0s9 leap 42.1: eth0 leap 42.2: eth0 tumbleweed: enp0s9 13.2 & TW go for 'predictable' names, leap for kernel names. Can't the community make up their mind? The mac address is not stored, so you can't relate interface names based on that. One way I see would be to re-do what 80-net-setup-link.rules does in yast to see which names the interfaces would have had. FWIW, I'd add net.ifnames=0/1 to my boot options and be done either way. Ludwig, was it a deliberate decision to go for classical interface names in Leap or did it just happen? Do we have a plan to consolidate either way? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1000732 http://bugzilla.suse.com/show_bug.cgi?id=1000732#c5 --- Comment #5 from rens groenewegen <rens.groenewegen@xs4all.nl> --- "The mac address is not stored, so you can't relate interface names based on that" sure. but we might consider to start doing that. that, or the slot/bus position of the ethercards and which names they might have according to that in ethX or enpXsX naming convention. indeed, it would be nice if we could make up our minds... cheers, Rens -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1000732 http://bugzilla.suse.com/show_bug.cgi?id=1000732#c6 Ludwig Nussel <lnussel@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(lnussel@suse.com) | --- Comment #6 from Ludwig Nussel <lnussel@suse.com> --- Known problem when updating from 13.2: https://doc.opensuse.org/release-notes/x86_64/openSUSE/Leap/42.1/#idm1402896... Yes, Leap aligns with SLE in that regard whereas TW follows upstream default systemd behavior. Nevertheless would be good if YaST could detect the situation if messed up and restore a working setup somehow. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1000732 http://bugzilla.suse.com/show_bug.cgi?id=1000732#c7 Ancor Gonzalez Sosa <ancor@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |ancor@suse.com Resolution|--- |FEATURE --- Comment #7 from Ancor Gonzalez Sosa <ancor@suse.com> --- Looks like this a documented behavior and the consequence of a conscious decision in the development of Leap, so not a YaST bug. Closing. Created a Fate entry to track the feature request suggested by Ludwig in the previous comment https://features.opensuse.org/321608 -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com