[Techtalk] initrd woes

Maria McKinley maria at shadlen.org
Mon Jan 26 11:20:30 UTC 2009

Maria McKinley wrote:
> It is an nfs mount, and that is where sbin is mounted at that point. Not 
> sure how or when it is moved to /. This is the last command that the 
> initrd script gives, so I guess init is suppose to take care of moving 
> /root. Init is a binary, and I haven't figured out how to troubleshoot 
> it yet, so I don't know if it is failing before it gets started or what. 
> The fail message I get on the console is:
> run-init: /sbin/init: No such file or directory
> [    11.607804] Kernel panic - not syncing: Attempted to kill init!
> Which would lead me to believe it is the exec run-init that fails rather 
> than init itself, but I am finding error messages are often misleading...
> thanks,
> maria


I have made significant progress! I can now start init. Turns out there 
are empty directories at the root level that need to be there that I 
hadn't copied to the nfsroot. Not sure if all of these were important, 
but adding emul, opt, selinux, and srv was what helped me. I'm not quite 
out of the woods yet, though. Still having nfs issues. I start up with 
the read only mounts that initrd made. This appears to be because 
rpc.statd is not running. And it appears that rpc.statd is not running 
because portmap is not running. But I don't know why portmap is failing. 
If I try to start it from the command line, it fails silently. It seems 
to be related to this bug, but this bug is suppose to be closed, and 
fixed in portmap 6.0-9. Of course, this is the version I am using...


Not sure how to proceed. Does anyone have experience with bugs that are 
suppose to be closed, but seem not really to be? Not even sure how or 
where to post for that.


More information about the Techtalk mailing list