[mythtv-users] FC5 Diskless boot woes

[Don Bossung]

> -----Original Message-----
> From: Don Bossung [mailto:donald at bossung.net]
> Sent: Friday, March 23, 2007 10:34 AM
> To: 'Discussion about mythtv'
> Subject: RE: [mythtv-users] FC5 Diskless boot woes
> 
> > Don Bossung wrote:
> > >
> > > I know that this may not be the appropriate forum for this problem,
> > > but I am hoping the someone here has had a similar problem as I am
> > > getting nowhere fast with other lists and google searches.
> > >
> > >
> > >
> > > Here we go.
> > >
> > >
> > >
> > > I am using
> > >
> > > FC5 2.6.20 x86_64
> > > the machine is running
> > > tftpserver 41-1.2.1
> > > dhcp 11:3.0.3-28
> > > system-config-netboot 1.40-1.FC5
> > >
> > > Mythtv 0.20-154
> > >
> > >
> > >
> > > The above is my backend and I am trying to setup a diskless
> > > workstation to use as a frontend.
> > >
> > >
> > >
> > > On the backend I:
> > >
> > > Followed the system-config-netboot docs at
> > > /usr/share/doc/system-config-netboot-1.40-1 using seciton 2 of this
> > > document steps 1-6
> > >
> > > For my setup machine I installed vmware server 1.02 on the above box
> > > and then setup a FC5 vm from cd iso with text install and no
> > > additional packages. Post install I used yum to update and then group
> > > install X Desktop Enviroment and KDE (K Desktop Environment) and
> > > MythtvFrontend.
> > >
> > > After following the docs above I booted the remote machine (Celeron D
> > > 2.88 1G RAM no drives) The pxe boot and kernel load goes fine until
> > > the following error occurs:
> > >
> > > RAMDIKSK: Compressed image found at block 0
> > > EXT2-fs: blocksize too small for device
> > > attempt to access beyond end of device
> > > ram0: rw=0, want=18446744056529682440, limit=44184
> > > isofs_fill_super: bread failed, dev=ram0, iso_blknum=17,
> > block=2147483648
> > > No filesystem could mount root, tried: ext2 iso9660
> > > Kernel panic - not synching: VFS: Unable to mount root fs on
> > > unknown-block(1,0)
> > >
> > > I don't get why so many blocks (18 pentillion???) are being requested.
> > > I don't think dhcp/tftp is the culprit as I am getting the kernel.
> > > Could it be that I used vmware as the base for creating the base image
> > > used by netboot on the remote machine?
> > >
> > >
> > >
> > > BTW, I used the vmlinuz and initrd.img from the Fedora Core 5 dvd and
> > > was able to load that with no problem using the same setup.
> > >
> > > Again, I know I am way off topic here, but if you have any ideas, let
> > > me know.
> > >
> >
> > I'll be the one to follow this up with a "me too" type of post, which
> > many on the list despise.  However, I think it is worth stating that you
> > are not alone in the scenario.
> >
> > I noticed the same problem and fought with it for a couple of nights,
> > playing with the "ramdisk_size=" and "ramdisk_blocksize=" kernel
> > arguments within the PXE config to adjust things.  I got past that
> > particular error with :
> >     append  initrd=MythTV4/initrd.img root=/dev/ram0 init=/disklessrc
> > NFSROOT=192.168.56.10:/export/diskless/mythtv4 ramdisk_size=21912
> > ETHERNET=eth0 SNAPSHOT=mythtv4  ramdisk_blocksize=1024 ip=dhcp
> >
> > While that got me past the error, I hit another show-stopper error a
> > couple of lines further down (where it is supposed to mount the
> > filesystems and execute /disklessrc and S01diskless), and that is where
> > I tossed in the towel.  I figured I would work on upgrading all of the
> > systems to FC6 and address the diskless setup again.
> >
> Thanks, it is good to see I am not the only one (wow did that come out
> wrong!) Did try the ramdisk size/blocksize with no luck. I think I will
> try the FC6 fixes listed in the diskless frontend howto and give it
> another go.

Thought I would update this with the solution I found. The first time I
built the setup machine in vmware, I used the default of LVM for swap/root.
I rebuilt the setup machine and re-rsynched everything and got to the login
prompt. I did add the ramdisk_blocksize=1024 as per the Diskless Frontend
howto in the wiki. When I tried to login, I got a message that flashed on
the screen then the login prompt would reappear. This happened regardless of
the user I tried to login as. Using Ctrl-S I was able to freeze the screen
and saw that is was an SELinux audit error message. I thought this was odd
as I had turned off SELinux on the setup machine before rsynching. So, I
went back to the pxe default config file and added selinux=0 to the append
line. All is good now and I can login to the diskless workstation.

Don