Scyld and fstab for Diskless slaves

[Chip Coldwell]

I know how to compile a kernel to NFS mount a root filesystem, and I
have set up diskless clusters using this method; I'm very curious how
the ramdisk root filesystem works.  Does the node tftp down a
compressed filesystem image and load it into the ramdisk analogously
to the root floppies of yore, except coming over tftp instead of from
a floppy?  Does this require custom kernel modifications, or is it a
functionality available in stock kernels?

Thanks,

Chip

-- 
Charles M. "Chip" Coldwell
"Turn on, log in, tune out"

On Fri, 15 Dec 2000, Daniel Ridge wrote:

> 
> On Fri, 15 Dec 2000, David Leunen wrote:
> 
> > But I have another prob: How to configure /etc/beowulf/fstab correctly
> > for diskless nodes? I didn't see anything about diskless in the doc.
> 
> For reference, the details of the slave fstab file are in the Beowulf
> install guide bundled with our distribution. To summarize:
> 
> Configuring ramdisks for Scyld Beowulf is straightforward.
> Comment or remove the '/' and 'swap' entries from /etc/beowulf/fstab.
> Create a line which looks like:
> 
> /dev/ram3	/	ext2	fs_size=65536	0 0
> 
> The magic bit is that we have added a mount option -- fs_size -- that
> tells us how large to make the disk.
> 
> This relates to another question asked recently in this forum:
> 	Why does our default kernel command line specify a ramdisk_size
> 	of 128M?
> 
> This should be pretty safe on boxes that don't even have 128M -- but
> you will get into trouble if you try to use the large ramdisks. If
> you have troubles installing Scyld Beowulf -- this isn't the problem.
> 
> Also, while we set up a beefy ramdisk during boot -- this space is
> reclaimed when the node finishes booting if you aren't doing a ramdisk
> root.
> 
> This, in turn, relates to another freqeuntly asked question:
> 	Why do my nodes show up as unavailable?
> 
> The reason is usually that the default fstab (/etc/beowulf/fstab) does
> not relate to the partitioning on your disks. Whatever the reason,
> the details of the node boot do not appear on the console. This record
> can be found in /var/log/beowulf/node.<nodenum> where <nodenum> is the
> node number that you seek.
> 
> Our upcoming contains a lot of effort directed towards reducing the size
> of ramdisk. This is to make the system work on smaller memory machines
> and to make diskless operation more palatable on medium memory machines. I
> have run our upcoming version on VMware instances with as little as 16M
> with a beowulf ram footprint of about 2M.
> 
> Because we can now do this, I expect that the next version of our product
> will ship with ramdisk root as the default.
> 
> Regards,
> 	Dan Ridge
> 	Scyld Computing Corporation
> 
> 
> 
> 
> _______________________________________________
> Beowulf mailing list
> Beowulf at beowulf.org
> http://www.beowulf.org/mailman/listinfo/beowulf
>