It's interesting [to me] Thomas
On 03/16/2013 04:32 PM, Thomas Lange wrote:
On Sat, 16 Mar 2013 13:27:52 -0400, George VerDuin <gfv2008-h...@yahoo.com>
said:
> missed? -- after the black screen fai continues to build for another
> 4-1/2 minutes before the fai-client automatically re-boots. NFS
> continues to deliver server content to the client that totals 79k
> packets. The client hard drive remains active until reboot -- just like
> a normal build. Even though the result is not 100% correct, fai is
> building something in the 5:50 min between power-on and re-boot.
Maybe FAI works perfectly.
slang: "Close but no cigar." -- it refers to the time just before a wife
gives birth...
I got lucky to find an already loaded USB with live ubuntu 10.10
notebook OS.
The fai-client HD was set up per FAIBASE, the first partition was
labelled bootable.
A lot was loaded into /usr, /var, and some other directories.
BUT -- /tmp was empty [except for lost+found] as was /home, and /boot.
The / directory had no links for vmlinuz or initrd.
Software sources are from wheezy.
From my preliminary exploration, I conclude that fai aborted some time
a) after the basic / filesystem was created and populated from fai-server
b) before the kernel and grub was loaded.
If I am to guess at a reason for failure without further evidence:
? --> fai-client has no access to internet to finish the install
from repo?
To check this, try to ssh login into the
install client. Then have a look at /tmp/fai where the log files are
stored before the client reboots.
It is understood that fai-client fails to re-boot for ssh exploration --
right?
I believe I saw somewhere that fai supports a ssh login during system
build time, uid=fai & pwd=fai.
And I suppose I could remove the "reboot" flag to give me more time to
explore?
Since /tmp/fai is now gone, I presume it was ram-disk content and
disappeared at reboot.
Is that exactly what you have in mind?
Cheers
Geo
