Re: OpenVMS?

Folks,

Would it be worthwhile to light up some buildfarm animals on OpenVMS?

http://www.openvms.org/stories.php?story=10/02/09/2319162

Cheers,
David.
--
David Fetter <david(at)fetter(dot)org> http://fetter.org/
Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter
Skype: davidfetter XMPP: david(dot)fetter(at)gmail(dot)com
iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

David Fetter wrote:
> Folks,
>
> Would it be worthwhile to light up some buildfarm animals on OpenVMS?
>
> http://www.openvms.org/stories.php?story=10/02/09/2319162
>
>
>

Sure, go for it.

cheers

andrew

David Fetter <david(at)fetter(dot)org> writes:
> Would it be worthwhile to light up some buildfarm animals on OpenVMS?

Have we ever even claimed to support VMS? I have no particular desire
to undertake a major new porting effort.

regards, tom lane

It could be interesting to see how big a porting effort it was ... ?

I'd say go for it and let's see what is involved ...

On Tue, 16 Feb 2010, Tom Lane wrote:

> David Fetter <david(at)fetter(dot)org> writes:
>> Would it be worthwhile to light up some buildfarm animals on OpenVMS?
>
> Have we ever even claimed to support VMS? I have no particular desire
> to undertake a major new porting effort.
>
> regards, tom lane
>
> --
> Sent via pgsql-hackers mailing list (pgsql-hackers(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-hackers
>

----
Marc G. Fournier Hub.Org Hosting Solutions S.A.
scrappy(at)hub(dot)org http://www.hub.org

Yahoo:yscrappy Skype: hub.org ICQ:7615664 MSN:scrappy(at)hub(dot)org

On Tue, Feb 16, 2010 at 08:11:15AM -0500, Andrew Dunstan wrote:
>
>
> David Fetter wrote:
> >Folks,
> >
> >Would it be worthwhile to light up some buildfarm animals on OpenVMS?
> >
> >http://www.openvms.org/stories.php?story=10/02/09/2319162
> >
>
> Sure, go for it.
>
> cheers
>
> andrew

Here's what I sent them:

First Name : David
Last Name : Fetter
Organization : PostgreSQL Global Development Group
E-mail Address : davidfetter(at)postgresql(dot)org
Products being ported : PostgreSQL

We'll also need

* Shell access from several accounts
* Git or cvs client
* Compiler tools
* Perl of a fairly recent vintage
* Outbound http access

Warm Regards,
David.

--
David Fetter <david(at)fetter(dot)org> http://fetter.org/
Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter
Skype: davidfetter XMPP: david(dot)fetter(at)gmail(dot)com
iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

On Tue, Feb 16, 2010 at 11:22 AM, David Fetter wrote:
>    * Shell access from several accounts
>    * Git or cvs client
>    * Compiler tools
>    * Perl of a fairly recent vintage
>    * Outbound http access

I had access to the HP testdrive before they closed it down (the Unix
servers were down in Sept 08, but the VMS cluster was running till Jan
2010). They blocked all outbound internet access to the testdrive
servers -- only telnet and ftp were allowed.

Count me in for the OpenVMS porting effort (but I guess I will apply
for an account seperately as I might port other things to OpenVMS in
the future). I believe the porting effort is larger than a new Unix
port but smaller than the Windows port, as most of the Unix and POSIX
functions and system calls are supported on OpenVMS.

MySQL has around 10 functions changed or written specifically for
OpenVMS, most of those are related to utime(), $UMASK & $UMASKDIR,
open(), and Unix pathnames. (I think utime() support was added a few
years ago to OpenVMS 7.3 & 8.0 -- so may be the code was added to
MySQL for earlier VMS versions.)

PostgreSQL uses fork(), which is not supported on OpenVMS. However,
the techniques used by the WIN32 version of internal_forkexec() in
src/backend/postmaster/postmaster.c give the VMS version a good
starting point.

Lastly, are we going to support ODS-2 disks?? And are we going to
require GNV installed for building and running PostgreSQL on OpenVMS??

Rayson

>
>    Warm Regards,
>    David.
>
> --
> David Fetter <david(at)fetter(dot)org> http://fetter.org/
> Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
> Skype: davidfetter      XMPP: david(dot)fetter(at)gmail(dot)com
> iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics
>
> Remember to vote!
> Consider donating to Postgres: http://www.postgresql.org/about/donate
>
> --
> Sent via pgsql-hackers mailing list (pgsql-hackers(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-hackers
>

On Tue, Feb 16, 2010 at 12:39:29PM -0500, Rayson Ho wrote:
> On Tue, Feb 16, 2010 at 11:22 AM, David Fetter wrote:
> >    * Shell access from several accounts
> >    * Git or cvs client
> >    * Compiler tools
> >    * Perl of a fairly recent vintage
> >    * Outbound http access
>
> I had access to the HP testdrive before they closed it down (the
> Unix servers were down in Sept 08, but the VMS cluster was running
> till Jan 2010). They blocked all outbound internet access to the
> testdrive servers -- only telnet and ftp were allowed.

Outbound http access is for the buildfarm, which is pretty important
to how we develop.

> Count me in for the OpenVMS porting effort (but I guess I will apply
> for an account seperately as I might port other things to OpenVMS in
> the future). I believe the porting effort is larger than a new Unix
> port but smaller than the Windows port, as most of the Unix and
> POSIX functions and system calls are supported on OpenVMS.

You clearly know vastly more than I do about this, and should lead
this effort :)

> Lastly, are we going to support ODS-2 disks?? And are we going to
> require GNV installed for building and running PostgreSQL on
> OpenVMS??

You tell us :)

Cheers,
David.
--
David Fetter <david(at)fetter(dot)org> http://fetter.org/
Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter
Skype: davidfetter XMPP: david(dot)fetter(at)gmail(dot)com
iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

> -----Original Message-----
> From: pgsql-hackers-owner(at)postgresql(dot)org [mailto:pgsql-hackers-
> owner(at)postgresql(dot)org] On Behalf Of Rayson Ho
> Sent: Tuesday, February 16, 2010 9:39 AM
> To: David Fetter
> Cc: Andrew Dunstan; PG Hackers
> Subject: Re: [HACKERS] OpenVMS?
>
> On Tue, Feb 16, 2010 at 11:22 AM, David Fetter wrote:
> >    * Shell access from several accounts
> >    * Git or cvs client
> >    * Compiler tools
> >    * Perl of a fairly recent vintage
> >    * Outbound http access
>
> I had access to the HP testdrive before they closed it down (the Unix
> servers were down in Sept 08, but the VMS cluster was running till Jan
> 2010). They blocked all outbound internet access to the testdrive
> servers -- only telnet and ftp were allowed.
>
> Count me in for the OpenVMS porting effort (but I guess I will apply
> for an account seperately as I might port other things to OpenVMS in
> the future). I believe the porting effort is larger than a new Unix
> port but smaller than the Windows port, as most of the Unix and POSIX
> functions and system calls are supported on OpenVMS.
>
> MySQL has around 10 functions changed or written specifically for
> OpenVMS, most of those are related to utime(), $UMASK & $UMASKDIR,
> open(), and Unix pathnames. (I think utime() support was added a few
> years ago to OpenVMS 7.3 & 8.0 -- so may be the code was added to
> MySQL for earlier VMS versions.)
>
> PostgreSQL uses fork(), which is not supported on OpenVMS. However,
> the techniques used by the WIN32 version of internal_forkexec() in
> src/backend/postmaster/postmaster.c give the VMS version a good
> starting point.

For PostgreSQL, you will probably want to use LIB$SPAWN() as a rough equivalent to CreateProcess() on Windows
http://www.sysworks.com.au/disk$vaxdocsep002/opsys/vmsos721/5932/5932pro_041.html

> Lastly, are we going to support ODS-2 disks?? And are we going to
> require GNV installed for building and running PostgreSQL on OpenVMS??

Dann Corbit wrote:
> > PostgreSQL uses fork(), which is not supported on OpenVMS. However,
> > the techniques used by the WIN32 version of internal_forkexec() in
> > src/backend/postmaster/postmaster.c give the VMS version a good
> > starting point.
>
> For PostgreSQL, you will probably want to use LIB$SPAWN() as a rough equivalent to CreateProcess() on Windows

Ah, LIB$SPAWN, that brings back memories.

> http://www.sysworks.com.au/disk$vaxdocsep002/opsys/vmsos721/5932/5932pro_041.html
>
> > Lastly, are we going to support ODS-2 disks?? And are we going to
> > require GNV installed for building and running PostgreSQL on OpenVMS??

I hate to pour cold water on this, but why is it worth adding support
for a platform that has such marginal usage.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +

On Tue, Feb 16, 2010 at 2:56 PM, Dann Corbit wrote:
> For PostgreSQL, you will probably want to use LIB$SPAWN() as a rough equivalent to CreateProcess() on Windows

We will need to support running PostgreSQL as a detached process, and
thus it won't have access to the DCL CLI. The implication is that
lib$spawn or system() will fail when used this way.

http://labs.hoffmanlabs.com/node/794

So we will likely need to use sys$creprc() in internal_forkexec().

Rayson

> http://www.sysworks.com.au/disk$vaxdocsep002/opsys/vmsos721/5932/5932pro_041.html
>
>> Lastly, are we going to support ODS-2 disks?? And are we going to
>> require GNV installed for building and running PostgreSQL on OpenVMS??
>
>

On Tue, 16 Feb 2010, Bruce Momjian wrote:

> I hate to pour cold water on this, but why is it worth adding support
> for a platform that has such marginal usage.

Because someone feels like dedicating their resources to it ... ?

----
Marc G. Fournier Hub.Org Hosting Solutions S.A.
scrappy(at)hub(dot)org http://www.hub.org

Yahoo:yscrappy Skype: hub.org ICQ:7615664 MSN:scrappy(at)hub(dot)org

Marc G. Fournier wrote:
> On Tue, 16 Feb 2010, Bruce Momjian wrote:
>
> > I hate to pour cold water on this, but why is it worth adding support
> > for a platform that has such marginal usage.
>
> Because someone feels like dedicating their resources to it ... ?

Well, there is going to be impact on the community too --- patches,
testing, etc. The community effort is small, but isn't zero.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +

Marc G. Fournier wrote:
> On Tue, 16 Feb 2010, Bruce Momjian wrote:
>
>> I hate to pour cold water on this, but why is it worth adding support
>> for a platform that has such marginal usage.
>
> Because someone feels like dedicating their resources to it ... ?

That's step one. Step two is community approval. Otherwise, anyone
with an idea would be a committer.

--
Andrew Chernow
eSilo, LLC
every bit counts
http://www.esilo.com/

On Tue, 2010-02-16 at 15:47 -0500, Andrew Chernow wrote:
> Marc G. Fournier wrote:
> > On Tue, 16 Feb 2010, Bruce Momjian wrote:
> >
> >> I hate to pour cold water on this, but why is it worth adding support
> >> for a platform that has such marginal usage.
> >
> > Because someone feels like dedicating their resources to it ... ?
>
> That's step one. Step two is community approval. Otherwise, anyone
> with an idea would be a committer.

With all respect to OpenVMS, I really can't help but ask, "why?". The
maintenance load over time will be large and I can't imagine this being
a wide use platform for us.

Joshua D. Drake

>
> --
> Andrew Chernow
> eSilo, LLC
> every bit counts
> http://www.esilo.com/
>

--
PostgreSQL.org Major Contributor
Command Prompt, Inc: http://www.commandprompt.com/ - 503.667.4564
Consulting, Training, Support, Custom Development, Engineering
Respect is earned, not gained through arbitrary and repetitive use or Mr. or Sir.

On Tue, 16 Feb 2010, Bruce Momjian wrote:

> Marc G. Fournier wrote:
>> On Tue, 16 Feb 2010, Bruce Momjian wrote:
>>
>>> I hate to pour cold water on this, but why is it worth adding support
>>> for a platform that has such marginal usage.
>>
>> Because someone feels like dedicating their resources to it ... ?
>
> Well, there is going to be impact on the community too --- patches,
> testing, etc. The community effort is small, but isn't zero.

But its totally worth it.

scrappy(at)hub(dot)org ("Marc G. Fournier") writes:
> On Tue, 16 Feb 2010, Bruce Momjian wrote:
>
>> I hate to pour cold water on this, but why is it worth adding
>> support for a platform that has such marginal usage.
>
> Because someone feels like dedicating their resources to it ... ?

But adding it in would require *some* degree of thought on the part of
committers as to "what might break VMS builds?"

If someone wants to make a branch to run on VMS, that would be a well
and fine thing.

It is quite likely that once we're on Git, that might be easily managed
by having a repo which feeds off "official" releases, modifying only in
those places where VMS-specific changes are required.

It might even turn out to be the case that the patches are sufficiently
small and undemanding that it would turn out to be easy to merge into
the "official" release, so as to make it an official platform.

I'd not want to assume that up front, though.
--
output = reverse("moc.liamg" "@" "enworbbc")
http://linuxdatabases.info/info/slony.html
"As long as each individual is facing the TV tube alone, formal
freedom poses no threat to privilege." --Noam Chomsky

rocrash(at)gmx(dot)de (Robert Doerfler) writes:
> On Tue, 16 Feb 2010, Bruce Momjian wrote:
>
>> Marc G. Fournier wrote:
>>> On Tue, 16 Feb 2010, Bruce Momjian wrote:
>>>
>>>> I hate to pour cold water on this, but why is it worth adding support
>>>> for a platform that has such marginal usage.
>>>
>>> Because someone feels like dedicating their resources to it ... ?
>>
>> Well, there is going to be impact on the community too --- patches,
>> testing, etc. The community effort is small, but isn't zero.
>
> But its totally worth it.

Do we have a patch yet?

If it's small and easy, then that points to you being right.

If ugly and hairy... then not so much...
--
(format nil "~S(at)~S" "cbbrowne" "gmail.com")
http://www3.sympatico.ca/cbbrowne/nonrdbms.html
"I think you ought to know I'm feeling very depressed"
-- Marvin the Paranoid Android

Bruce Momjian <bruce(at)momjian(dot)us> writes:
> Marc G. Fournier wrote:
>> On Tue, 16 Feb 2010, Bruce Momjian wrote:
>>> I hate to pour cold water on this, but why is it worth adding support
>>> for a platform that has such marginal usage.
>>
>> Because someone feels like dedicating their resources to it ... ?

> Well, there is going to be impact on the community too --- patches,
> testing, etc. The community effort is small, but isn't zero.

No, I suspect the community effort would be *large*. VMS is
sufficiently unlike Unix that this port would probably be akin to the
Windows port in terms of invasiveness, fragility, and general need
for everyone to bend over backwards for it. The discussion about
fork substitutes should give you some idea of what we'd be in for.

I think the odds of getting the community to support such a port
are not easily distinguishable from zero, and I agree with Bruce's
desire to dissuade anyone from pouring effort down the drain.

regards, tom lane

All,

In my 12 years on the PostgreSQL project, this is the 2nd time, ever,
I've heard a question about OpenVMS support. The previous time was in 2003.

Maybe there's an untapped community out there, but personally I think
we'd find more users on z/OS than on OpenVMS.

--Josh Berkus

Josh Berkus <josh(at)agliodbs(dot)com> writes:
> In my 12 years on the PostgreSQL project, this is the 2nd time, ever,
> I've heard a question about OpenVMS support. The previous time was in 2003.

Well, a search of our archives for "OpenVMS" finds a few more, but it
still looks like about one request a year.

If we could have a port for an amount of effort roughly commensurate
with that level of interest, I'd be all for it. But I'm afraid the
work would be far out of proportion to that.

regards, tom lane

On Tue, Feb 16, 2010 at 1:02 PM, David Fetter wrote:
> You clearly know vastly more than I do about this, and should lead this effort :)

I have been working on OpenVMS on and off for a few years, but I am
sure a lot of people are more qualified than I am :)

I guess I will create a page in the PostgreSQL wiki about the port,
and put the ToDOs list in there.

To others: Given that Ingres was ported to OpenVMS, and given that
many other opensource apps have been ported to OpenVMS without a lot
of effort, I believe we should at least try before giving up OpenVMS.

Rayson

P.S. Ingres has a number of VMS users:

http://community.ingres.com/forum/dba-forum/653-will-ingres-ported-openvms-itanium.html

http://www.groupsrv.com/computers/post-3095363.html

>> Lastly, are we going to support ODS-2 disks?? And are we going to
>> require GNV installed for building and running PostgreSQL on
>> OpenVMS??
>
> You tell us :)
>
> Cheers,
> David.
> --
> David Fetter <david(at)fetter(dot)org> http://fetter.org/
> Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
> Skype: davidfetter      XMPP: david(dot)fetter(at)gmail(dot)com
> iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics
>
> Remember to vote!
> Consider donating to Postgres: http://www.postgresql.org/about/donate
>