I just moved my system (which was on a single hard drive) to several software raid volumes (one for /, one for /home, etc). I made an image of the previous root filesystem, and dd'd it to the new RAID volume. Then, I recompiled the kernel to ensure that software raid was compiled in (and not just a module), and fiddled with grub a bit to get the system to boot.

Long story short, everything's peachy except for one crucial element: hald (on runlevel default) always segfaults when I boot. It also segfaults when I try to start it from /etc/init.d. Here's the error I'm getting:

This is a problem because without hal, I have no mouse or keyboard in X!

My system is up-to-date, and I've tried re-emerging hal and hal-info.

Any ideas?

Hello Secant,

I'd suggest to use the strace command and try to find out what's going on when hald starts and when it is stopped.

Markus

Thanks for the advice. I ran strace and piped the output to the attached file. Unfortunately, I don't really know what to look for. I noticed that there are several error codes (-1) being returned (ENOENT and ECHILD) errors while 'stat'ing files from /var/lib/init.d but I don't know if that's normal or what.

I noticed the script can stat some files in /var/lib/init.d but not others (notably hald itself). What could this mean? Other daemons aren't starting up properly?

If anybody knows how to interpret this information, I'd really appreciate some help.

Attached Files

haldlog.txt (155.3 KB, 18 views)

Problem solved. It turns out that hal-0.5.12-r8 (the most current stable version) can't deal with the root partition being on an mdraid device. Luckily for me, a newer version is available in portage (hal-0.5.13-r2) that doesn't segfault. I unmasked it and a quick emerge later I was on my way.