Solaris

provided that they are of the same h/w architecture with same network drivers etc, technically it should work.

Hi,

Firstly, your question isnt too clear here, so not 100% sure what your asking, so will take a guess.

Zones - You have two types of Network configurations, shared / exclusive. During the configuration of shared NIC, you will defined card / ip / netmask via zonecfg, so once the zone is up, the card / network nic will be up and running. Exclusive is different and all you do is define the NIC that will be `exclusively` used within the zone. Once up, you will then need to configure the card like any other normal environment, i.e. /etc/hosts, /etc/hostname.nxge1, /etc/defaultrouter etc etc etc. Then this will be automatic on every reboot on the zone.

Exclusive IP
dladm to see what devices are there
zonecfg:myzone> set ip-type=exclusive
zonecfg:myzone> add net
zonecfg:myzone:net> set physical=e1000g1
zonecfg:myzone:net> end

Shared IP

add net
set address=XX.30.188.XX/24
set physical=nxge4
end

Next point, copying of the /etz/zones/index file. Ok, yes Ive done this, but you are not really meant to do this due to the uid given to the zone. ALso, its the zonename.xml which is also important. Ive copied zones, mainly for DR purposes. I tend to build / define a zone, get it up and running and in a state that its working / configured. I then shut it down, detach it, dump the configuration to a file, and copy over the file / dump of the environment to the target DR server. I then copy the dump into the desired location, create the new zone usin g the zonecfg dump configuration file, but renaming it accordingly, then attach the zone.

The attach part gets over the zoneadm -z xxxxx install part, where the new zone is built. By attaching it, it tells the zoneadm that there's already an environment there, to attach to.

Here, command wise, is what I would do.

1) Backup zone you want to copy, detach zone first
global# zoneadm -z zfszone1 detach / remember to attach afterwards
]2) create copy
global# tar / cpio the directory of the source zone to target location
The source zone is just a directory with all the files in, so just get it copied over to the new location, whether on the same host under a new name, or a dfifferent hot
3) attach zone / create new zone
zonecfg -z zfszone2
zonecfg:zfszone2> create -a /zone_roots/zfszone2
zonecfg:zfszone2> exit
4) attach and boot zone
global# zoneadm -z zfszone2 attach
If on a different host, then the attach part will check versions / patches etc. You will need to alter the ip / hostname as it’s a copy


Hope the above helps.

you just need to modify your /etc/zones/zonename.xml file.

zonecfg -z zonename
add net
set physical=ce0
set address=x.x.x.x

or you can just vi the file in /etc/zones and you should be fine as well.

Hello Friends,
Today I got the downtime to check whether on a switchover the newly added interfaces would be available on the other VCS node as well. My tests failed. I will try to sum up what exactly had I done:

==================================================================
1) On local zone (primary cluster node), I added the interfaces using "zonecfg -z zonename"
2) I then plumbed each interfaces using the commands:
ifconfig <interface:name> plumb
ifconfig <interface:name> zone <zone name> inet <ip addr> netmask <netmask> up

3) Copied the zone xml file to the secondary cluster node. Also ensured that the permissions, ownership and size of the xml file are same on both the nodes.

4) Did the switchover
hagrp -switch <zone name/package> -to <other node>

5) "Ifconfig -a" did not show the newly added interfaces.
==================================================================

Any help to let me know where I went wrong would be highly appreciated.

Thanks&Regards,
Mahive.

Hi

With cluster, all network interfaces will have to be exculsive, i.e. the global zones should present the underlying NICs to the zone, which will not be shared with any other zones, and be exclusive to that one.

So, for each cluster zone, you should be setting set ip-type=exclusive Then, with in the zone, you do the usual /etc/hosts / etc/hostname.nxge1 etc etc.

In the global zone, go into /etx/zones, and ensure both of the zonename.xml files for the cluster zones, have the network address lines, mine look like :-

<zone name="svx-nemesis-dr" zonepath="/zones/roots/svx-xcxxxx-dr" autoboot="true" brand="solaris8"bootargs="-m verbose" scheduling-class="FSS" ip-type="exclusive">
<network address="" physical="nxge1"/>
<network address="" physical="nxge2"/>

I think you need to step out of the thinking of zones, and see each zone as a separate server. Once you have presented each environment with its own NIC, then put they are zones to one side, and configure the environments like you would do for a clustered pair.

I would imagine a private network address, or etc etc.

Do a google aswell.

SBK