Configuring two systems with multiple Network Interface Cards to support Transaction Logging between them for 'hot standby' (FFR) purposes.

mvBase supports using two systems to provide a ‘hot standby’ for a ‘live’ system. This feature is known as Fast Fault Recovery (FFR) and is a set of procedures rather than specific pieces of software. FFR depends on inter Transaction Logging (TL) to move file updates from the live system to the standby system. This ensures that the two copies of the database are kept in synch.

On any system, even without FFR, Transaction Logging can be performed and the data written to a tape device, thus creating a set of log tapes. These could be used to restore transactions should a catastrophic system problem occur. The restore could take place on the same system, once the problem has been resolved, or on a spare system, after the full database has been File loaded from the most recent File Save. FFR allows virtually real-time, automatic, (requiring no operator intervention) updates of the standby system to occur.

On mvBase the tape device can be a virtual one as well as the traditional physical hardware tape drives. One of the two types of virtual tape device is the Virtual Tape Link (VTL); the other is the Virtual Tape File (VTF). The Virtual Tape Link allows a communications link between the two systems to function as a pair of ‘tape drives’. The VTL on the live system is used to log transactions to, and at the same time, the VTL on the standby system is used to read the transactions from so that they can then be applied to the standby database.

When Transaction Logging is performed over a Virtual Tape Link, then depending upon the number of file updates being generated on the live system, it could generate a significant amount of network traffic. If the network interface card(s) and physical network are being used for other purposes (normal file and print serving, or more likely, to connect the mvBase clients to the mvBase server) then overall user response may suffer.

If this is the case, then a solution is to place an additional Network Interface Card (NIC) in the both systems connected directly to each other over a separate crossover network cable (see below).

The steps below serve as a guide to set this configuration up. It does not describe the creation of Virtual Tapes Links, or the use of the Transaction Logger, which are documented in the user manuals.

1. Install 2 Network Interface Cards in each system and ensure that they do not have hardware resource conflicts (IRQ’s base port addresses)
2. If the systems are normally called LIVE and STANDBY, to correspond to the (for example) 192.100.100.x subnet IP addresses, then the secondary NIC’s IP addresses (192.100.101.x) will be resolved from different names (say LIVE_FFR and STANDBY_FFR). Thus each physical system will have two names which resolve into two different IP addresses, each on a different subnet and connecting through a different NIC.
3. Configure the first card on each system to be on your normal network subnet, by allocating an appropriate IP address and associated netmask, and setup its default gateway. e.g.

   LIVE Nic 1:    IP Address:   192.100.100.1

   Netmask:    255.255.255.0

   STANDBY Nic 1:    IP Address:   192.100.100.2

   Netmask:    255.255.255.0

4. Configure the second card on each system into a different subnet e.g.

   LIVE_FFR Nic 1:    IP Address:   192.100.101.1

   Netmask:    255.255.255.0

   STANDBY_FFR Nic 1:    IP Address:   192.100.101.2

   Netmask:    255.255.255.0

5. Install a crossover Ethernet cable between the two secondary cards (ie the 192.100.101.x cards) or connect the cards to their own dedicated network segment by having a single hub with only these two cards connected to it.
6. Add an entry into the hosts file which is located in (%systemroot%\system32\drivers\etc\hosts) on both systems as follows:

   On LIVE:

      192.100.101.2      STANDBY_FFR

   On STANDBY:

      192.100.101.1      LIVE_FFR

7. Reboot both servers and test using ping that you are in fact routing the network traffic through the right cards.
8. On LIVE, do a ping 192.100.101.2 (or a ping STANDBY_FFR) and watch the individual ‘link active’ lights on the hub to see if the hub port connected to the secondary NIC’s on LIVE and STANDBY are being used.
9. If the two systems are directly connected with a crossover cable, instead of via a hub, then the NT Network Monitor utility should be used to examine traffic on the subnet (192.100.101.x).

Finally, check that that you can still access both systems via their primary NIC’s from elsewhere on the main network.