On 1 February 2015 11:04:46 CET, Bob McKay <rimsnucse@xxxxxxxxx> wrote: >Hi Anders; apologies for missing your post earlier. You and Brian both >suggested dkms, but the problem with dkms is that if it fails, the >system will just fail to reboot - which means physically going to the >lab to fix things, and would be a disaster if I am overseas. So I think >sticking with manual building is necessary. Thanks for the point about >devel; but it isn’t the cause of the problem - I always wait to rebuild >until kernel headers and devel for the new kernel are available (the >local mirrors are sometimes a bit slow to update, so I’m very aware of >the risk of mismatched latest versions). By manually rebuilding, I can >see that the build failed while I’m still in a running system, and I >can usually repair and reboot to the preceding system when it fails. >Since I’m usually doing this remotely, that’s a really important plus. Perhaps instruct grub not to use the new kernel, and let dkms switch grub default when build succeds? >Anders Blomdell said... > >Use appropriate scripts in /etc/kernel/postinst.d or >/etc/kernel/postinst.d/dkms >and kernel modules can be rebuilt right after a new kernel is installed >(of course you need kernel-devel). > >/Anders >_______________________________________________ >Yum mailing list >Yum@xxxxxxxxxxxxxxxxx >http://lists.baseurl.org/mailman/listinfo/yum -- Sent from my Android device with K-9 Mail. Please excuse my brevity. _______________________________________________ Yum mailing list Yum@xxxxxxxxxxxxxxxxx http://lists.baseurl.org/mailman/listinfo/yum
