[systemd-devel] Systemd segv during boot

[Lennart Poettering]

On Mon, 03.09.12 09:59, Nicolas Aguirre (aguirre.nicolas at gmail.com) wrote:

> Hi
> 
> I'm using Angstrom and systemd (v44) on an ARM custom board. (TI OMAP3
> Cortex a8)
> 
> Sometimes (around 1 time on 100) I get a systemd segv during the boot,
> otherwise the system boots fine and everything is OK.
> I'm using a RO file sysytem on a UBI/UBIFS partition.
> 
> I dont know how to go further and see why a segv happened (how to get a
> backtrace for example), You maybe  will have ideas about how to debug this
> issue ?

If your root disk is writable you should automatically get a coredump in
the root dir. On SIGSEGV PID 1 will fork and create that coredump in the
child, so expect this coredump to belong to another PID than 1.

> Caught <SEGV>, core dump failed.
> Freezing execution.

This suggests that your root dir was not writable hence the coredumo
could not be generated.

Lennart

-- 
Lennart Poettering - Red Hat, Inc.