[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference iosg::all-in-1_v30

Title:*OLD* ALL-IN-1 (tm) Support Conference
Notice:Closed - See Note 4331.l to move to IOSG::ALL-IN-1
Moderator:IOSG::PYE
Created:Thu Jan 30 1992
Last Modified:Tue Jan 23 1996
Last Successful Update:Fri Jun 06 1997
Number of topics:4343
Total number of notes:18308

3331.0. "Question about ALL-IN-1 on ALPHA" by THELAB::ASBRIDGE (Vince Asbridge) Mon Sep 27 1993 17:01

    I have done a quick search of this notes file, and am admitidly an
    ALL-IN-1 illiterate, but I haven't been able to find any reference
    to an ALL-IN-1 kit for Alpha AXP machines.
    
    Please tell me I'm looking in the wrong places,  and an Alpha based
    ALL-IN-1 kit is alive and well.
    
    I got a bit of a scare when the following note was sent to me by
    mail...
    
    From:   MKOTS4::REDZIN::DCOX "from MKO1, DTN 264-0969  24-Sep-1993
    1530"
    24-SEP-1993 15:29:27.13
    To:     SHISHOV BECKLEY
    CC:     BCLARK,DCOX
    Subj:   Sable and operating systems
    
    Andrei, Stew
    
    If/since an Alpha machine cannot be attached to a cluster that has
    ALL-IN-1 running on ANY node....
    
    And if/since nobody in SW Engineering is tasked to correct the
    problem...
    
    And if/since ALL-IN-1 is on, arguably, 90% of clusters.....
    
    And if/since some number of Sable sales are expected to be into
    existing clusters....
    
    Have you all given much thought as to what that will do to Sable
    penetration?
        
    Dave

Editor's Note: Spelling of ALL-IN-1 (tm) corrected by conference moderator.

T.RTitleUserPersonal
Name
DateLines
3331.2It's happeningIOSG::ECHRISTIEEileen ChristieTue Sep 28 1993 09:303
Dave,
	I'd be happy to give you lots more information, either via A1INFO or if
you mail me. We are very committed to Alpha AXP.
3331.3Says who?IOSG::PYEGraham - ALL-IN-1 Sorcerer's ApprenticeTue Oct 05 1993 14:089
    I know of *ABSOLUTELY NO REASON* (yes, I am shouting :-) ) why a mixed
    architecture cluster shouldn't support any version of ALL-IN-1 running
    on the VAX nodes only.
    
    If anyone knows differently, please tell me *NOW*!!
    
    Graham Pye (ALL-IN-1 Engineering, Installation Supervisor)
    
    PS Shortly it'll run on the AXP nodes too!
3331.4ALL-IN-1 works fine...IOSG::PYEGraham - ALL-IN-1 Sorcerer's ApprenticeMon Nov 01 1993 11:194
    We've now discovered this is a Message Router, not ALL-IN-1
    restriction. We're looking into the full story.
    
    Graham
3331.5Apologies to Message Router!IOSG::PYEGraham - ALL-IN-1 Sorcerer's ApprenticeWed Nov 03 1993 08:478
    I'm told by several Message Router Experts that they can't see *ANY
    REASON* (I'm shouting on their advice!) why Message Router shouldn't
    work either.
    
    We're now going to ask the original reporter of this problem for more
    details.
    
    Graham
3331.6Fixed in MR 3.3?FORTY2::ASHMail Interchange Group, ReadingThu Nov 04 1993 09:4310
In lieu of a statement from 'those who know', I'll add my 2p to this debate - 
my understanding is that there is a problem with running MR on mixed clusters, 
specifically with DDS, which uses the cluster alias and assumes it can use all 
nodes in the cluster.

However, this assumption/restriction is likely to be removed with MR V3.3, and 
at that time it will be perfectly possible to run the MR processes on the 
VAXes only of a mixed cluster.

grahame
3331.7Listen to Graham!FORTY2::ASHMail Interchange Group, ReadingThu Nov 04 1993 16:565
oops. Please ignore the previous reply.

.5 has the definitive statement.

grahame
3331.8Don't allow the alias to connect to AlphaIOSG::TALLETTGimmee an Alpha colour notebook...Tue Nov 09 1993 20:056
    
    	Presumably you could set up the Alpha's so that they don't field
    	cluster alias requests? Or wouldn't that help?
    
    Regards,
    Paul
3331.9CSOA1::LENNIGDave (N8JCX), MIG, @CYOWed Nov 10 1993 12:5720
    Exactimundo!  There is nothing IOS/MR unique in this situation.
    
    If you had a VAX-only cluster (Phase IV DECnet) with an alias where 
    some services were not available on a particular VAX, this is your
    only recourse (exec alias inc dis). The same applies to AXP nodes.
    
    (Note that if you don't do this, it isn't the end of the world wrt
    MR - both TS and DDS will simply retry, and eventually the PhIV alias
    link dispatch mechanism will select a 'valid' MR cluster node.)
    
    Note that if the cluster is using Phase V DECnet, the alias link
    dispatch mechanism has gotten smarter, and will only direct links 
    to nodes that support the target/requested 'object'. Hence you can
    allow all nodes full use of the alias with no ill effects.
    
    Dave
    
    PS - MR V3.3, MRG V3.3, and MRX V3.3 all support DECnet/OSI (subject to
    availability of node synonyms) and have been submitted to SSB. You'll 
    have to talk to ALL-IN-1 product mgmt about IOS support on DECnet/OSI.