Date: Thu, 19 Jan 2017 07:36:52 +0330 From: mokhi <mokhi64@gmail.com> To: leres@ee.lbl.gov Cc: freebsd-ports@freebsd.org, julian@freebsd.org Subject: Re: poudriere: devel/llvm39: runaway_process Message-ID: <CAByVWPUbZaVeM62cP025ZvVcQjOb=_Lb6OD1PKBqQUUVWdkA7A@mail.gmail.com> In-Reply-To: <3dad3665-c635-00c9-3638-a72ba6bf774b@freebsd.org> References: <303419bf-7502-3f79-ce7a-a0bb720a9fd5@ee.lbl.gov> <3dad3665-c635-00c9-3638-a72ba6bf774b@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 1/19/17, Craig Leres <leres@ee.lbl.gov> wrote: > My poudriere build servers have been unable to build devel/llvm39 ever > since some ports changed from llvm37 (January 17th?) The error is > runaway_process and this happens during the package phase. It looks like > pkg-static spins for one hour at which point poudriere gives up on it. > Yesterday I raised max_execution_time for package on one build server to > 2 hours which did not help. My three servers are all at > 10.3-RELEASE-p16 and I'm attempting to build packages for the same versio= n. > > I'm still able to build llvm37 ok. Since it takes ~2 hours for the > failure to occur I'm hoping for ideas on how to diagnose this. > Well, I'm not sure about the error (mine builds fine), but as a helping note, maybe passing `-i` option to `poudriere testport` helps you go into the jail after the fail, then no need to rebuild all things after fail =F0=9F=98=81=F0=9F=98=84=F0=9F=98=8A simply change and ma= ke. However sometimes in some cases this approach fails, but generally I can say it works (for reducing test/fix time).
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAByVWPUbZaVeM62cP025ZvVcQjOb=_Lb6OD1PKBqQUUVWdkA7A>