Enabling >1 NIC active

[Dave Gudewicz]

I see LAN failover in new versions of VMS. 1 NIC active, the other waiting for something bad to happen, then it takes over.

Can the standby NIC be made to do something other than waiting? Like spreading the in and outbound network load, etc.

I saw the load broker thing, but it needs to write to a DNS server and that we cannot do.

[Andreas Fassl]

Hi Dave,

the feature is called FailSafe IP.
In
http://h71000.www7.hp.com/doc/732final/6526/6526pro_005.html#failsafe_chap

you find:
"The failSAFE service monitors an interface and takes appropriate action upon detecting interface failure or recovery. failSAFE IP provides IP address redundancy by requiring the same IP address to be configured on multiple interfaces. Only one instance of each IP address is active at any time; the other duplicate IP addresses are in standby mode."

failSafe IP is a redundancy solution - not a performance enhancement feature.

Loadbroker:
This is only interesting, if you've got many clients accessing your hosts.

Example:
client: whois vmsfarm.company.com
Loadbroker identifies a system by configurable match criterias and returns an IP address

This is explained in detail at:
http://h71000.www7.hp.com/doc/732final/6526/6526pro_016.html#lbroker_chap

TCP/IP Services provides two methods for balancing work load using DNS:

* Load sharing using the default DNS method of round-robin scheduling.
* Load balancing using the TCP/IP Services load broker. Load broker is a configurable, calculated, load-balancing mechanism for distributing the work load among DNS cluster members.

If you can't write to the DNS server, talk to you network guys about the concept of subdomains. Probably it is an approach to your problem to introduce a sub-domain
host1.vms-systems.company.com to be administrated by a local vms DNS-server.

Regards

Andreas

[Dave Gudewicz]

Sounds like the subdomain, local VMS DNS server idea is something to explore.

[Keith Parris]

LAN Failover and failSAFE IP are two different things.

LAN Failover works down at the LAN adapter (MAC address) level, and works for all protocols. failSAFE IP works only for IP, but can fail over not only to another adapter on the same node, but even to an adapter on a different node in the cluster.

There's a good article on failSAFE IP (which includes a good table comparing and contrasting this with LAN Failover) in an article entitled "Configuring TCP/IP for High Availability" by Matt Muggeridge in Volume 2 of the OpenVMS Technical Journal at http://h71000.www7.hp.com/openvms/journal/v2/index.html

The best description of LAN Failover in the documentation seems to be in a section entitled "Understanding LAN Failover" at the end of chapter 10 in the 'System Manager's Manual, Volume 2: Tuning, Monitoring, and Complex Systems' at http://h71000.www7.hp.com/doc/732FINAL/aa-pv5nh-tk/aa-pv5nh-tk.PDF

If you'd like a backup LAN adapter in a LAN Failover set to do some work, you could allow cluster traffic to run over it (PEDRIVER uses all LAN adapters by default, unless you disable its use of an adapter or change relative priorities), or point other protocols to the underlying LAN adapter device instead of the LLc0: LAN Failover Set device (but then they'd not benefit from LAN Failover, of course, and would have to do failover at a higher level).

[Dave Gudewicz]

Thanks for the links Keith.

After reading, the failSAFE IP idea is a good one and is replacing IP Cluster Alias for HA. It also provides outbound connection load balancing.

I need to read a bit more on the idea of providing inbound load balancing, which is a performance consideration. The default round robin method seems OK, but can that be made to fit in or co-exist with failSAFE IP I wonder.