| Mauricio,
Not really, at least not both active at the same time. I presume
that you want some type of fail-over capability? If so, then the
best bet is to use the same approach; that is, use DECsafe, with PCM
running on both. If it's not for failover, then you should be able to
get by with one, providing that your network is fast enough.
If you do decide to go with 2, then you will have to shutdown PCM
on one system and manually start up the other system.
Dave
|
| >We're participating on a BID where the prospect is asking for both
>a local console AND a remote console capability. Our configuration is
>a dual AlphaServer 2100 4/200 with DECsafe Available Server. My
>question: is it possible to implement such a "beast" with Polycenter
>Console Manager? In other words, can we have TWO stations running
>Polycenter Console Manager (one local, one remote, both via "reverse
>LAT/TELNET") and being the console for the SAME pair of DECsafe systems?
>Any hint will be most welcomed!
The answer to your question is both yes and no. The "no" part is due to
the fact that a decserver port cannot have more than one active
host-initated LAT connection or more than one telnet connection at the
same time. To clarify we'll use an example:
We have two PCM engines called PCM1 and PCM2. We have a datacenter
with several systems that we want to manage. All the systems in the
datacenter are hooked to a decserver called PCMDS1. The configuration
database on both PCM1 and PCM2 is identical. We start Console Manager
on PCM1 and it establishes communciations with the decserver ports and
thus assumes control of the consoles of the datacenter systems that
are attached to it. We now start console manager on PCM2 but it cannot
connect to the decserver ports because PCM1 already has done so. This
same scenario existed in VCS by the way. Note also it will make no
difference as to what protocol we use - LAT or TELNET - the results
will be the same.
The "yes" answer to your question comes about because of the
transportable nature of the X-windowing system. Let's assume that
our example scenario has changed such that PCM2 is several miles away
in another city. The operators that have access to PCM2 would be able
to connect to PCM1 and SET DISPLAY back to PCM2. This would allow them
to send the C3 back to PCM2 and provide the most convenient way to
manage all the systems in the configuration. Note that everything is
still going through PCM1 though. You can also send all the X-windows
based action routines to either PCM1 or PCM2 as well. Note also that
when you are set host into the active PCM engine that you could also
run and of the character-celL interfaces such as MONITOR and CONNECT.
One of the up and coming features within PCM will be active failover.
Using our scenario above with PCM1 and PCM2 with active failover, if PCM1
was in control and failed due to a system crash, PCM2 would immediately
assume control of the systems in the datacenter.
Finally, some sites have attempted to create a totally duplicate
config right down to the decservers. They then install a Y cable
between the servers such that there are two physical paths to the same
console. I personally do not recommend this as there are electrical
problems as well usage problems, i.e., what happens when two operators
are connected to the same node at the same time and start typing
commands at the same time. You wind up with unintelligible garabage
on the console! The electrical problems are enough or a reason not to
do it. I have found that we can receive characters easialy enough but
when something is sent to the console it doesn't make it. I've seen
everything from corrupt characters to nothing happening at all! Someone
may have designed an "intelligent" cable that overcomes these problems
but I still don't recommend it.
Good luck!
Dan
|