Alas, that plan didn't work either.

I can't "exclusively" use in a local zone an interface plumbed in the global 
zone (i.e. not e1000g0):

[EMAIL PROTECTED] /]# zoneadm -z t1 reboot
WARNING: skipping network interface 'e1000g0' which is used in the global zone.
zoneadm: zone 't1': WARNING: skipping network interface 'e1000g0' which is used 
in the global zone.: Not owner

I can not use the vni interfaces for the similar reason - the global zone won't 
have an alias there so it can't be a router. Even worse, if the vni1 interface 
is not plumbed, the local zone can't attach to it:

/etc/zones/t1.xml :
...
<zone name="t1" zonepath="/zones/t1" autoboot="false" ip-type="exclusive
">
  <inherited-pkg-dir directory="/lib"/>
  <inherited-pkg-dir directory="/platform"/>
  <inherited-pkg-dir directory="/sbin"/>
  <inherited-pkg-dir directory="/usr"/>
  <network address="" physical="vni1"/>
</zone>

[EMAIL PROTECTED] /]# zoneadm -z t1 reboot
zoneadm: zone 't1': WARNING: unable to hold network interface 'vni1'.: Invalid 
argument

If I do "plumb" it in a global zone (not even "up" it), it's "used" and the 
local zone can't attach to it either:

[EMAIL PROTECTED] /]# ifconfig vni1 plumb

[EMAIL PROTECTED] /]# zoneadm -z t1 reboot
WARNING: skipping network interface 'vni1' which is used in the global zone.
zoneadm: zone 't1': WARNING: skipping network interface 'vni1' which is used in 
the global zone.: Not owner

So I believe now that the exclusive IP stacks are of limited if any usability 
in this scenario (unless I did wrong in the tests above). That is, I could use 
them for up to a couple of local zones if we had the extra connections actually 
wired to a switch...

Now, when is a new generation of networking projects (Crossbow, Nemo, 
Clearview) going to deliver from "just around the corner" phase to the 
"download a package/patch, install on Solaris 10/OpenSolaris Nevada, try out 
and cheer" phase?
 
 
This message posted from opensolaris.org
_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to