Thanks.

I just got the ZFS stuff working.  I was just confused on the device notation 
and the need to run "/usr/sbin/disks" to get the device nodes created after I 
added a new drive.

In my case, c0t0d0, c0t2d0, c0t3d0 were the SATA devices that I needed to 
specify to "zpool create ..." , but I never saw those exact files under 
/dev/dsk, just the variants with the "p0/p1/p2/p3" or "s0/s1/s2/...", so I 
didn't trust that it would work.

The UFS EIDE boot disk did have a c1d1s0 entry under /dev/dsk, which further 
confused me about why the SATA disks weren't showing up like that.

BTW, after running zpool, I do see c0t0d0, etc.  So that problem was mostly a 
case of my own Solaris ignorance.

As far as the e1000g, in my case the ethernet chip is embedded on the 
motherboard, and uses Intel's CSA port on the 865G northbridge to avoid 
congesting the PCI bus / ICH5 southbridge with network traffic.  I would expect 
that variation to be fairly common (i.e. well tested) in higher end 
motherboards since CSA was supposed to be a real win for gigabit ethernet, but 
maybe I'm wrong.

Perhaps I should try moving the SATA controller to a different PCI slot just in 
case there is an IRQ issue of some sort (gawd, sounds like the old ISA bus 
days).

And my hostname *is* screwed up (i.e. assigned as "unknown") due to an apparent 
expectation that the DHCP server would be tell it what the hostname should be, 
rather than having the hostname hardcoded.  So my /etc/hosts has this entry at 
the end:

192.168.1.120   unknown # Added by DHCP

And my /etc/hostname.e1000g0 is empty.

The command-line installation process never prompted me for a hostname, so 
maybe I just need to tweak one file or so, but I haven't research how DHCP is 
managed by Solaris yet.  I should probably just disable DHCP, since that 
probably isn't the best idea for a NAS.

Thanks again for your ideas, and any others that this might spur.

--Mark
 
 
This message posted from opensolaris.org
_______________________________________________
opensolaris-help mailing list
[email protected]

Reply via email to