vswitch failover

[paabhi]

Hi Team,

We are setting IVMs for our dev environment. The Version using is 4.1.We have 2 hosts on which 6 guests are running and the host servers are clustered.

We are having 2 data lans on each host.Currently vswitch is created on primary lan and asssigned to all VMs running on the host.The second data lan is used as failover lan in SG configuration.

If the primary lan fails, the vswitch will not failover to second data lan.

Is there any option we can have redundancy for the vswitch ? APA configuration on the host is not allowed as the hosts are clustered.

Abhilash.

[Eric SAUBIGNAC]

Bonsoir Abhilash,

APA configuration on the host IS supported.

Check page 40 in this document : http://bizsupport1.austin.hp.com/bc/docs/support/SupportManual/c02015679/c02015679.pdf

or § 7.2.12, page 66 in the HPVM 4.2 releases notes : http://bizsupport1.austin.hp.com/bc/docs/support/SupportManual/c02123997/c02123997.pdf

What is not supported is a vswitch backed by a VLAN interface on the host (see previous doc)

I have configured most of my clients with MC/ServiceGuard (11.18, 11.19) and Integrity Virtual Machines (4.0, 4.1 and trying today with 4.2 ;-) as Serviceguard Packages. At the host level the vswitch are always backed by an APA interface. And it do works ...

It seems that the vswitch monitor is also able to move a vswitch from a failed nic to a standby one. Don't know how, so you will have to RTFM.

Regards

Eric

[Eric SAUBIGNAC]

More informations about vswitch monitor : check script /sbin/init.d/vswitchmon

In a few words, this script checks /var/adm/syslog/syslog.log to track failed lan messages from MC/ServiceGuard and in such a situation move automatically a vswitch from the failed lan to a stanby lan configured in MC/ServiceGuard.

I find this process is not so well integrated than APA. So I strongly suggest using APA.

You should also check http://forums13.itrc.hp.com/service/forums/helptips.do?#34

Eric