Comment 12 for bug 961482

Revision history for this message
Jim Leinweber (jiml-mail) wrote :

Wow, this bug is getting a lot of attention! Thanks, everyone.

Joseph Salisbury asked:
   When the system hangs, can you see if you can access some of the other ttys?

Sorry, it hangs too early for VTY switching. Specifically, the last two messages from around 1.47 seconds into boot are tagged: "start_kernel" and "gs_change". If the more specific address info would be useful I can transcribe the hex.

Joseph also asks:
   ... we could run a kernel bisect ...

Yes, I could install and try a series of other kernels. It's a work laptop, and I don't have time to build them from source, but I can run boot experiments with kernels someone else is compiling to try and pin this down.

In response to GNAServicesInc, yes bug 961823 sounds like the same one as mine. It's the same symptoms, anyway. The "recovery boot" options are noisier than the default boot and more informative. What I did to use apport-bug was first run "dpkg-query -l "linux*" to identify the full name of the kernel package I wanted to complain about.

Brad Figg notes:
    ... there is a newer version of the development kernel ...

I ran update-manager again, installed the 3.2.0-20 kernel, and it has the same symptoms.

Joseph Salisbury asks:
    ... would it be possible for you to test the latest upstream kernel?
I installed 3.3.0-030300.201203182135 and got a kernel panic at 0.32s into boot instead. Different, but not better.

Does anyone know if the "SerialConsoleHowto" document results in not just grub but also the kernel logging out a serial port? Maybe I could scrounge up a null modem cable and use kermit on another PC to capture the boot messages?