[Info-vax] Porting to Linux instead of x86-64 VMS, was: Re: 3rd party SATA SSD usage on RX2660+P400 in a production environment?

[Terry Kennedy]

On Tuesday, July 2, 2019 at 1:22:47 PM UTC-4, Simon Clubley wrote:
> On 2019-07-02, Stephen Hoffman <seaohveh at hoffmanlabs.invalid> wrote:
> >
> > And in two or three years, we'll either be starting to replace these 
> > existing servers with x86-64 and ports to that, or we'll be starting up 
> > the Linux ports again.
> 
> Are people actually talking about doing the latter ?
> 
> If so, is this due to the current VMS on x86-64 timescales or is
> it due to some lack of information coming out of VSI ?

Some portion of the installed base is running very old VMS (for example, see today's post about 7.3-2, presumably on a VAX). There will be a fair bit of work getting applications written for that environment to compile using later HPaqital toolchains on later VMS versions, assuming the source code is still available and someone knows how to actually build it.

If the source has been lost and a rewrite is in order, then the existing functionality needs to be carefully evaluated to determine how deeply (if it all) it relies on VMS features, as opposed to just running on VMS.

Given the commoditization of x86 hardware, "throwing more hardware at the problem" is less of an issue than it was in the VAX days (or Alpha, or Itanium). Vendor licensing policies (per-core / per-socket / per-phase-of-the-moon) still come into play, and in fact carry a greater weight now, since the hardware is so inexpensive.

I haven't seen any public disclosure of VSI pricing (regardless of hardware) and I don't know what their plans are regarding sockets / cores / etc. However, someone with a [potential] checkbook in hand can probably get at least a general idea of VSI's thinking.

If I were in the situation of the VAX 7.3-2 user, I would do the following:

1) On a test system (could be an emulator, but definitely not on the production system) verify that the existing sources can be built with the existing tools and create working executables.

2) Get a test Alpha (probably a used actual Alpha), whatever interim licenses are needed from VSI, etc. and try compiling / testing the sources. Any work to clean up the sources here should be directly useful on x86. This will give you an idea of how much work will be needed to get running with current versions of VMS and tools. But don't do too much work until you have...

3) An idea of what your licensing costs from VSI, etc. on x86 are likely to be, as well as a general sense of what hardware might be supported - if you're an all-Dell shop, for example, there may be issues getting a HP system in the door if the initial support will only be on selected HP models.