Guys, Thanks for the help so far, now comes the more interesting questions ...
Piggybacking off of some work being done to minimize Solaris for embedded use, I have a version of Solaris 10 U2 with ZFS functionality with a disk footprint of about 60MB. Creating a miniroot based upon this image, it can be compressed to under 30MB. Currently, I load this image onto a USB keyring and boot from the USB device running the Solaris miniroot out of RAM. Note: The USB key ring is a hideously slow device, but for the sake of this proof of concept it works fine. In addition, some more packages will need to be added later on (i.e. NFS, Samba?) which will increase the footprint.
My ultimate goal here would be to demonstrate a network storage appliance using ZFS, where the OS is effectively stateless, or as stateless as possible. ZFS goes a long way in assisting here since, for example, mount and nfs share information can be managed by ZFS. But I suppose it's not as stateless as I thought. Upon booting from USB device into memory, I can do a `zpool create poo1 c1d0', but a subsequent reboot does not remember this work. Doing a `zpool list' yields 'no pools available'. So the question is, what sort of state is required between reboots for ZFS?
Regards, -- Jim C _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
