https://bugzilla.novell.com/show_bug.cgi?id=255677 ------- Comment #2 from j.sendler@os.bvnk.org 2007-03-21 16:18 MST ------- (In reply to comment #1)

Looks like update problem. Please attach y2logs.

OK, attachment follows in a tick... To save You lots of unnecessary work, before that here's one brief note about how this happened: During the online update stage of a new install via internet, when it appears for the first time, there comes a warning like: "Do not update your kernel if you're just installing OpenSuSE. Do it later." Under that warning there were both an "OK" and a "Cancel" button, so I hit "Cancel" to skip the kernel update this time. Unfortunally, the new kernel was installed in spite of that, and in a way that, obviously, the modules dir of the old kernel was replaced by that of the new one, but the image had not been changed yet. So upon reboot after that, it was still kernel 2.6.18.2 running and hence the modules were no longer found, because they were now under the new 2.6.18.8 modules directory. "uname -a" showed that the .2 kernel was running. OK, after "depmod" failed, I went into the /lib dir, and since there was only the .8 subdirectory there, I first had the impression that it must have been some change backwards from .8 to .2, but when I look at that from today, it was the other way round. Because of the minor subversion change, the symlinking worked and helped me to get the system up, to then run "/sbin/lilo" and "depmod" with the new kernel manually. == Summary: I think that's not a problem with the update in itself, it's only that, once started, there was no way to stop the kernel from being replaced in this process. - I found that a bit irritating. It should be relatively easy to let the update function detect whether it is being called from an installation environment or not, and then to inhibit any kernel changes in that former situation. However, depending on the situation (and, security considerations), it might be an even better idea to first get the latest kernel, and only after that update all other things. In either case, when a kernel update is retrieved, the updater should do the "complete" job (i. e. make it the default kernel with all dependencies resolved), be it by keeping the old kernel as a safe option when a boot loader is present or not - or the updater should only copy the files and offer some dialogue (i. e. to _keep_ the modules dir of the old kernel until it becomes redundant). Well, Cyril, thanks for Your patience, attachments ahead. Cheers, Jack. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=255677 j.sendler@os.bvnk.org changed: What |Removed |Added ---------------------------------------------------------------------------- Component|Basesystem |Update Problems Summary|depmod not working! (solved)|[renamed] keep modules dir for replaced kernels | |after update -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=255677 jsuchome@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jsuchome@novell.com, sh@novell.com AssignedTo|jsuchome@novell.com |schubi@novell.com ------- Comment #5 from jsuchome@novell.com 2007-03-27 03:17 MST ------- Michael: I'm not sure if we have such popup - definitely not as a part of online-update code. I think all warning texts should be part of kernel patch (installation notification/EULA). If user canceled the popup, it looks like either UI didn't react properly, or solver didn't remove the kernel package. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.