Conference iosg::all-in-1_v30

    Although in theory it is possible to run multiple systems of the same
    version on the same system - We have about four or five development
    versions running at the same time, there isn't an installation option
    to set it up, and it wouldn't be supported.
    
    There would be quite a penalty in disc space and memory use for all the
    images, FLCs, TXLs etc. too, and unless you set it up very carefully,
    you might need to give the users privileges. Multi-lingual systems are
    a problem too if that's relevant.
    
    Is that enough words? Tell them I wrote it and I said they shouldn't do
    it :-)
    
    Graham

    Further to my original note, and Grahams reply, I asked the following
    question - anyone out there got any comments?
    
    Cheers,
    
    Andy
    
    
From:	WARLOC::DORAN        17-MAR-1992 16:23:36.59
To:	IOSG::PYE,DORAN
CC:	
Subj:	ALL-IN-1 question

Graham,

Thanks for your reply to my question (note 256) in the ALL-IN-1 conference. 
Further to that, I have another question which you may be able to help with 
(it is related to the original...).

If the customer had 5 remote sites, would it be feasible to build a 5 node 
cluster and move the ALL-IN-1 systems from each remote site onto the 
cluster. Each node in the cluster could then be booted to offer ALL-IN-1 as 
a service, but the service is unique to each node.

ie - NODEA boots ALL-IN-1 from DISK1 
NODEB boots ALL-IN-1 from DISK2
etc.

User from remote site A connects to NODEA for an ALL-IN-1 service
User from remote site B connects to NODEB for an ALL-IN-1 service
etc.

Obviously, each node would effectively be a system in its own right - 
meaning upgrades to ALL-IN-1 would need to be carried out 5 times on the 
cluster. However, the saving would be in that one system could be managed 
from one location with one system management group.

Over time, the separate ALL-IN-1 'systems' could be merged, giving a 
homogeneous Cluster.

Would this be

a) supported

b) a 'good' idea?

Cheers,

Andy

P.S. forgive me not using ALL-IN-1, but the power to the building died, and 
our office system hasn't recovered yet!
    
    
    
From:	IOSG::IOSG::MRGATE::"A1::PYE" 17-MAR-1992 18:25:15.31
To:	WARNUT::DORAN
CC:	
Subj:	RE: ALL-IN-1 question

From:	NAME: Graham Pye                    
	FUNC: IOSG                            
	TEL: DTN 830 3451                     <PYE AT A1 at IOSG at REO>
To:	NAME: DORAN <DORAN@WARNUT@MRGATE@IOSG@REO>

Andy,

You'd do better to ask this sort of question in the notesfile to get a wider 
range of answers. Feel free to post this reply too.

I think it's supported to have separate systems like this, but I don't know if 
you'd need to have separate SYSUAFs etc. Not sure what the Message Router etc. 
implications would be either.

You'd have to be very careful how you set it up to make sure that all the batch 
jobs ran on the right node.

Although you could manage it from one location, it would be five times as much 
management. You could presumably do it just as easily distributed using VCS or 
similar.

We don't have a lot of practical experience in Engineering, we just build it 
and the people in the field derive the optimum ways to run it!

If you do this, make sure that you keep all the account names, DLs etc. 
uniquely named, and set the systems up with prefereably different SDAFs, but 
definitely different active ranges of shared areas.

I'm not going to answer your question about whether it's a good idea!!!

Good Luck,

Graham