Is it possible to replicate from one instance to (
except for the underlying hardware)an identical instance. I've read (Carlton Doe) that there can be issues because the logical logs can call for a new extents. Is this just a space issue ? Or, are there more complex trouble spots for using this method?

We have a SUN E450, Ultra10 and an Ultra2 all running Solaris 8 with
the same version of Informix and tools. The replication needs to be done asynchronously.

If you want to replicate the whole instance (this means
every database inside this instance and every table
inside these databases) you can use HDR.

The two machines need to be the same, that means they
need to have the same processor architecture.
However they don't need to have the same number
of processors.

HDR is stable and is used at some big Informix customers.
You can perform asynchronous and synchronous replication
with HDR.


The other option is using ER (Enterprise Replication).
With ER you can replicate only a subset of the tables.
They can be replicated to more than one server and
also an update anywhere configuration is possible.
The participating servers can consist of different hardware
architectures.
However, ER is more difficult to setup and maintain and the
replication is always asynchronously.

Yes, I am aware of HDR and ER; that is why I asked the question if it is possible to use the logical logs; since, that is the basic method that
HDR & ER use.
One could use triggers for replication but there are performance issues.

Question: Do logical logs make any calls that are specific to the underlying hardware ?

The setup is 1 identical database on 2 different SUN hardware that are both running the same version of Solaris 8.

If the two SUN machines are binary compatible
you can use HDR.

If for example one of these machines is a sparc and the other
is an intel pc with solaris, you can't use HDR but you can
use ER.

Eric,

I appreciate your taking the time to reply to this thread; it is nice to hear from someone who has extensive knowledge with Informix.

I find myself even more interested now. Is it possible to send the logical logs to a file and then ftp the file to an off-site twin machine( the machine is a fraternal twin since everything is identical except one machine is a SUN E450 and the other is a SUN Ultra 2) and perform a restore on the off-site fraternal twin?
The reason for these convoluted machinations is security.
If the CIA can be hacked, my conclusion is that trying to protect an internal network is a monetary hamster wheel; you can never get off; there are always new hacks. It is just a large ongoing cost for the sake of convenience, at least for a company of our size. And, we consider our data sensitive.
From my frugal understanding, I think when the logical logs are used for a restore, the logs make system calls ; these calls should be the same for 2 or more machines that are running SPARC/Solaris 8.
Could you please enlighten ?

Yes, you can save the logical logs to a file and transfer
this file to the secondary machine.

This hot-standby-setup has been used in the past by many big SAP installations.

You must first restore a physical level-0-archive on the second
machine and from this point on you can rollforward thru the
logical logs, i.e. keeping the second machine always in
rollforward recovery.

But you need some scripts which will automate this process
for you. This might not be as easy as it sounds.

You should also consider using VPN technology (virtual private
networks) for protecting your netword and then using HDR
for replicating your Informix instance. This sounds more reliable
to me.

Use storage replication software such as Veritas volume replicator to replicate storage only ,and use the external restore(onbar -r -e ) feature , It'll leave server in Fast Recovery mode and bring logical logs(as files or tapes ) to logical log restore and never restore L 0 before logical log restore