Thread: ppc64le support in 9.3 branch?
So I see somebody at 2ndQ has set up a bunch of ppc64le buildfarm
members, which I applaud. But they're all failing on the 9.3 branch,
because we lack support for that architecture in that branch.
Does anyone have the stomach for trying to add such support? The minimum
requirement would be to back-patch 9.4's config.guess and config.sub,
because that's where the builds are falling over right now. I wouldn't be
too afraid of that, but what is not clear is what portability issues might
be lurking beyond that. I could not find any specific mention of ppc64
in the git changelogs, but that doesn't mean there weren't any other 9.4
fixes that might need to be back-ported.
It's hard to justify putting in very much effort to add new-platform
support in a branch that's scheduled to die in six months, so I'm not
sure what to do. Should we just tell 2ndQ not to bother running those
animals on 9.3? Or should we make at least a bit of effort towards
making it work?
The compromise I'm inclined to offer is to see what happens if we
back-patch 9.4's config.guess and config.sub. If that makes these
animals go green, and doesn't break any others, we'll call it good.
Otherwise, we revert that change and say we're not putting any
additional effort into it.
regards, tom lane
Hi, On 2018-03-23 14:54:46 -0400, Tom Lane wrote: > So I see somebody at 2ndQ has set up a bunch of ppc64le buildfarm > members, which I applaud. But they're all failing on the 9.3 branch, > because we lack support for that architecture in that branch. > > Does anyone have the stomach for trying to add such support? The minimum > requirement would be to back-patch 9.4's config.guess and config.sub, > because that's where the builds are falling over right now. I wouldn't be > too afraid of that, but what is not clear is what portability issues might > be lurking beyond that. I could not find any specific mention of ppc64 > in the git changelogs, but that doesn't mean there weren't any other 9.4 > fixes that might need to be back-ported. > > It's hard to justify putting in very much effort to add new-platform > support in a branch that's scheduled to die in six months, so I'm not > sure what to do. Should we just tell 2ndQ not to bother running those > animals on 9.3? Or should we make at least a bit of effort towards > making it work? > The compromise I'm inclined to offer is to see what happens if we > back-patch 9.4's config.guess and config.sub. If that makes these > animals go green, and doesn't break any others, we'll call it good. > Otherwise, we revert that change and say we're not putting any > additional effort into it. I'm inclined to just ask them to stop running the animals on that branch. There are no pre-existing users on 9.3 ppc64le, and new customers hopefully won't move to 9.3. ISTM backpatching is riskier than just changing a bunch of buildfarm configurations. Greetings, Andres Freund
On Fri, Mar 23, 2018 at 7:58 PM, Andres Freund <andres@anarazel.de> wrote:
Hi,
On 2018-03-23 14:54:46 -0400, Tom Lane wrote:
> So I see somebody at 2ndQ has set up a bunch of ppc64le buildfarm
> members, which I applaud. But they're all failing on the 9.3 branch,
> because we lack support for that architecture in that branch.
>
> Does anyone have the stomach for trying to add such support? The minimum
> requirement would be to back-patch 9.4's config.guess and config.sub,
> because that's where the builds are falling over right now. I wouldn't be
> too afraid of that, but what is not clear is what portability issues might
> be lurking beyond that. I could not find any specific mention of ppc64
> in the git changelogs, but that doesn't mean there weren't any other 9.4
> fixes that might need to be back-ported.
>
> It's hard to justify putting in very much effort to add new-platform
> support in a branch that's scheduled to die in six months, so I'm not
> sure what to do. Should we just tell 2ndQ not to bother running those
> animals on 9.3? Or should we make at least a bit of effort towards
> making it work?
> The compromise I'm inclined to offer is to see what happens if we
> back-patch 9.4's config.guess and config.sub. If that makes these
> animals go green, and doesn't break any others, we'll call it good.
> Otherwise, we revert that change and say we're not putting any
> additional effort into it.
I'm inclined to just ask them to stop running the animals on that
branch. There are no pre-existing users on 9.3 ppc64le, and new
customers hopefully won't move to 9.3. ISTM backpatching is riskier than
just changing a bunch of buildfarm configurations.
+1 for dropping it.
On Fri, Mar 23, 2018 at 08:01:53PM +0100, Magnus Hagander wrote: > On Fri, Mar 23, 2018 at 7:58 PM, Andres Freund <andres@anarazel.de> wrote: >> I'm inclined to just ask them to stop running the animals on that >> branch. There are no pre-existing users on 9.3 ppc64le, and new >> customers hopefully won't move to 9.3. ISTM backpatching is riskier than >> just changing a bunch of buildfarm configurations. > > +1 for dropping it. +1. If something like that were to happen for 10 or 9.6 knowing that they still have four years to go, that could be debated, but just for 6 months there is little benefit. -- Michael
Attachment
On Mon, Mar 26, 2018 at 4:50 PM, Michael Paquier <michael@paquier.xyz> wrote: > On Fri, Mar 23, 2018 at 08:01:53PM +0100, Magnus Hagander wrote: >> On Fri, Mar 23, 2018 at 7:58 PM, Andres Freund <andres@anarazel.de> wrote: >>> I'm inclined to just ask them to stop running the animals on that >>> branch. There are no pre-existing users on 9.3 ppc64le, and new >>> customers hopefully won't move to 9.3. ISTM backpatching is riskier than >>> just changing a bunch of buildfarm configurations. >> >> +1 for dropping it. > > +1. If something like that were to happen for 10 or 9.6 knowing that > they still have four years to go, that could be debated, but just for 6 > months there is little benefit. I am in discussions with Mark, he's going to disable the animals from building 9.3. (by setting branches_to_build to 'HEAD_PLUS_LATEST4' instead of 'ALL'). cheers andrew -- Andrew Dunstan https://www.2ndQuadrant.com PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
Andrew Dunstan wrote: > I am in discussions with Mark, he's going to disable the animals from > building 9.3. (by setting branches_to_build to 'HEAD_PLUS_LATEST4' > instead of 'ALL'). So once we desupport 9.3, they'll stop building 9.4? :-) -- Álvaro Herrera https://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
On Tue, Mar 27, 2018 at 10:57 AM, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote: > Andrew Dunstan wrote: > >> I am in discussions with Mark, he's going to disable the animals from >> building 9.3. (by setting branches_to_build to 'HEAD_PLUS_LATEST4' >> instead of 'ALL'). > > So once we desupport 9.3, they'll stop building 9.4? :-) > Then we'll set it back to ALL :-) cheers andrew -- Andrew Dunstan https://www.2ndQuadrant.com PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
On Tue, Mar 27, 2018 at 11:09:37AM +1030, Andrew Dunstan wrote: > On Tue, Mar 27, 2018 at 10:57 AM, Alvaro Herrera > <alvherre@alvh.no-ip.org> wrote: > > Andrew Dunstan wrote: > > > >> I am in discussions with Mark, he's going to disable the animals from > >> building 9.3. (by setting branches_to_build to 'HEAD_PLUS_LATEST4' > >> instead of 'ALL'). > > > > So once we desupport 9.3, they'll stop building 9.4? :-) > > > > > Then we'll set it back to ALL :-) I'll put it on my calendar. :) Regards, Mark -- Mark Wong http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, RemoteDBA, Training & Services
Re: Tom Lane 2018-03-23 <1219.1521831286@sss.pgh.pa.us> > The compromise I'm inclined to offer is to see what happens if we > back-patch 9.4's config.guess and config.sub. If that makes these > animals go green, and doesn't break any others, we'll call it good. > Otherwise, we revert that change and say we're not putting any > additional effort into it. Fwiw, apt.postgresql.org has been building+regression testing 9.3 on ppc64el on various Debian+Ubuntu releases for more than a year now. The only requirement is updating config.guess/sub. Christoph
Christoph Berg <myon@debian.org> writes:
> Re: Tom Lane 2018-03-23 <1219.1521831286@sss.pgh.pa.us>
>> The compromise I'm inclined to offer is to see what happens if we
>> back-patch 9.4's config.guess and config.sub. If that makes these
>> animals go green, and doesn't break any others, we'll call it good.
>> Otherwise, we revert that change and say we're not putting any
>> additional effort into it.
> Fwiw, apt.postgresql.org has been building+regression testing 9.3 on
> ppc64el on various Debian+Ubuntu releases for more than a year now.
> The only requirement is updating config.guess/sub.
FWIW, I'm still in favor of that back-patch. But I was outvoted
before, and I dunno if this new info will change anyone's mind.
regards, tom lane