On Wed, Apr 1, 2015 at 10:15 AM, Martin Langhoff <martin.langhoff@xxxxxxxxx> wrote: > On Wed, Apr 1, 2015 at 11:42 AM, Martin Langhoff > <martin.langhoff@xxxxxxxxx> wrote: >> See below. Perhaps the newer kernel (in latest F21) has regressed in >> handling some kinds of errors during mount, or the dracut/systemd >> mounting process is less resilient than mounting under a fully booted >> system? > > This is getting even more interesting. > > Under 3.17.4-301.fc21.x86 from LiveUSB, I could mount, even repair the disk. > > Since the repair, the on-disk latest kernel (3.18.9-200.fc21) tries to > boot, but dracut/systemd time out on mounting sysroot after waiting > for quite a while. I don't get a dracut shell anymore so the failure > mode has changed. I may try to set a breakpoint to force a shell. > > I do have an earlier F21 kernel on disk-- 3.18.7-200.fc21 -- and this > boots the system without a glitch. After a complete boot with > 3.18.7.200, clean shutdown and booting into 3.18.9-200 is still > broken, same failure mode. > > Will try to capture some info from a dracut breakpoint (I'll try > mount). At this point this really looks like a regression. Yeah I don't know what's going on, but with a new file system, and disabled i915 to avoid crashes, and thus no crashes since the new fs was created, I get boot failure with 3.19.3 but not 3.19.2, and I can't figure out why. I get the systemd cylon eye with 5 services pending so I can't actually tell which one it's hung up on, but one of them is looking for the fs volume UUID and apparently can't find it which is completely bogus. -- Chris Murphy -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
