[Info-vax] OpenVMS servers and clusters as a cloud service

[Stephen Hoffman]

On 2018-01-11 19:11:28 +0000, DaveFroble said:

> Stephen Hoffman wrote:
>> On 2018-01-11 05:20:02 +0000, DaveFroble said:
>> 
>>> In our applications, we have a message manager, and all background 
>>> tasks get their assignments through the message manager.  When we want 
>>> them down, we just send a command to the message manager, and it 
>>> directs all active background tasks to shut down gracefully.  Rather 
>>> simple.  Works very well.
>>> 
>>> Even if some of the things you mention were available, application 
>>> designers would still have to figure out how to use them, and to 
>>> implement that usage.
>> 
>> No, your approach does not work well.
> 
> Well, gee Steve, some of this stuff has been working, WELL, since the 
> 1970s. Started on RSTS/E.  So, how long does something need to work, 
> WELL, before it meets your approval?

>> Sure, it works.  It certainly works for your particular current needs.  
>>  Well enough.

Good on you.   What you have works well enough for your needs.   But 
I'd wager that there are capabilities that — if you thought about it — 
that you'd like to have added, and there are things that could work 
better.    From having implemented and updated these home-built 
communications systems, remote access via TLS, maybe?  The ability to 
send commands to manage multiple servers in parallel?   The ability to 
send a command to check and ready updates, or to check installed 
versions?   The ability to enable and manage and configure to receive 
status and error notifications via web push notifications?   Donno.  
Spit-balling here.

Since we're all doing parts of these implementations, we all end up 
with what's "good enough" and maybe with a few of the "good" bits added 
if the there was some extra time in a schedule somewhere.

I don't expect you'd immediately migrate to a newer approach.   Most 
folks reasonably won't.   But you'll probably consider migrating when 
working on some major update in the future, and you will seriously 
consider adopt the newer approach for your newer designs.  This as the 
capabilities equal or exceed what you're presently working with.    
This rather than going through the effort involved in the existing 
management communications, and the the effort involved in implementing 
checkpointing capabilities using the existing low-level APIs.

I'm pushing for a higher baseline for OpenVMS.   Specific and better 
reasons for folks to choose it.   Of raising all the boats in the 
harbor.

This built-on-first-principles reinventing-the-same-wheel-in-every-app 
stuff just gets.... tedious.   Slogging through getting TLS support 
working (as I'm doing yet again) gets... tedious.



-- 
Pure Personal Opinion | HoffmanLabs LLC