Re: Searching for: Fast windows buildfarm animal - Mailing list pgsql-hackers

From Andres Freund
Subject Re: Searching for: Fast windows buildfarm animal
Date
Msg-id 20180424033514.kg2t7ccxkdfdeiqz@alap3.anarazel.de
Whole thread Raw
In response to Re: Searching for: Fast windows buildfarm animal  (Stephen Frost <sfrost@snowman.net>)
List pgsql-hackers
On 2018-04-20 19:55:26 -0400, Stephen Frost wrote:
> Greetings,
> 
> * Andres Freund (andres@anarazel.de) wrote:
> > It's common that half the buildfarm has reported back before a single
> > windows buildfarm animal reports. And if they report a failure one often
> > has to wait for hours for the next run.
> 
> Yes, that's been rather annoying.
> 
> > It'd be awesome if somebody could set up a windows animal that runs
> > frequently (i.e. checks for build needed every minute or five) and is
> > fast enough to not take ages to finish a build.
> 
> Done.
> 
> 'dory' was approved today which checks for new commits every 5 minutes
> and, as an added bonus, builds with a somewhat more recent version of
> Visual Studio (2015) and a recent Windows Server version (2016) as
> compared to the other Windows systems.  While it's still not quite as
> fast as our faster buildfarm animals, it's at least able to report in
> within 20-30 minutes or so, as compared to hours.
> 
> We'll probably continue to tweak the system a bit, but it's at least up
> and running now for HEAD, REL_10 and REL9_6 (older branches don't work
> easily with VS2015, as discussed elsewhere).
> 
> If anyone has any specific questions about it or suggestions, please
> feel free to reach out.

Thanks!

Greetings,

Andres Freund


pgsql-hackers by date:

Previous
From: Amit Kapila
Date:
Subject: Re: Toast issues with OldestXmin going backwards
Next
From: Michael Paquier
Date:
Subject: Re: wal_consistency_checking reports an inconsistency on masterbranch