[pkg-go] Bug#852959: Bug#852959: prometheus FTBFS on armhf: github.com/prometheus/prometheus/storage/local fails
On 31/01/17 13:01, Julien Cristau wrote: > On 01/28/2017 08:24 PM, Martín Ferrari wrote: >> On 28/01/17 16:10, Adrian Bunk wrote: >> >>> AFAIK the buildds are building one package at a time, >> >> Uhm, don't know about that, but I many times I have experienced test >> failures which are somehow timing-related on some buildds, failures I >> can only reproduce with very high CPU contention. >> > Did you set DEB_BUILD_OPTIONS in your attempts on abel? You can see the > parallel settings used on buildds in the logs. In the end, I started getting the failures on abel too, so there was no need for that. -- Martín Ferrari (Tincho) ___ Pkg-go-maintainers mailing list Pkg-go-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-go-maintainers
[pkg-go] Bug#852959: Bug#852959: prometheus FTBFS on armhf: github.com/prometheus/prometheus/storage/local fails
On 01/28/2017 08:24 PM, Martín Ferrari wrote: > On 28/01/17 16:10, Adrian Bunk wrote: > >> AFAIK the buildds are building one package at a time, > > Uhm, don't know about that, but I many times I have experienced test > failures which are somehow timing-related on some buildds, failures I > can only reproduce with very high CPU contention. > Did you set DEB_BUILD_OPTIONS in your attempts on abel? You can see the parallel settings used on buildds in the logs. Cheers, Julien ___ Pkg-go-maintainers mailing list Pkg-go-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-go-maintainers
[pkg-go] Bug#852959: Bug#852959: prometheus FTBFS on armhf: github.com/prometheus/prometheus/storage/local fails
On 28/01/17 16:24, Martín Ferrari wrote: >> and the machines where prometheus failed have 4 GB RAM. > > abel too... Now it fails on abel too.. I guess it was always too close to the limit, and anything else running killed it. >> "2680815616 in use", that is 2.6 GB. >> >> You have only 3 GB address space in userspace, >> and I'd guess that is the limit you hit. > > possibly. > >> Is this test expected to use huge amounts of memory? > > I will re-run the tests with /usr/bin/time on abel and see how much > memory do they take. Time says the maximum resident size was 2184076 KB, it is a bit less on amd64 (with tests passing). Now I don't know what to do.. If I upload a fix (disabling the test), it will be too late for Stretch. -- Martín Ferrari (Tincho) ___ Pkg-go-maintainers mailing list Pkg-go-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-go-maintainers
[pkg-go] Bug#852959: Bug#852959: prometheus FTBFS on armhf: github.com/prometheus/prometheus/storage/local fails
On 28/01/17 16:10, Adrian Bunk wrote: > AFAIK the buildds are building one package at a time, Uhm, don't know about that, but I many times I have experienced test failures which are somehow timing-related on some buildds, failures I can only reproduce with very high CPU contention. > and the machines where prometheus failed have 4 GB RAM. abel too... > "2680815616 in use", that is 2.6 GB. > > You have only 3 GB address space in userspace, > and I'd guess that is the limit you hit. possibly. > Is this test expected to use huge amounts of memory? I will re-run the tests with /usr/bin/time on abel and see how much memory do they take. -- Martín Ferrari (Tincho) ___ Pkg-go-maintainers mailing list Pkg-go-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-go-maintainers
[pkg-go] Bug#852959: Bug#852959: prometheus FTBFS on armhf: github.com/prometheus/prometheus/storage/local fails
On Sat, Jan 28, 2017 at 03:59:45PM -0300, Martín Ferrari wrote: > On 28/01/17 11:58, Adrian Bunk wrote: > > Source: prometheus > > Version: 1.5.0+ds-1 > > Severity: serious > > > > https://buildd.debian.org/status/fetch.php?pkg=prometheus=armhf=1.5.0%2Bds-1=1485604675=0 > > As it can be seen in the log: > > runtime: out of memory: cannot allocate 4194304-byte block (2680815616 > in use) > fatal error: out of memory > > Yesterday I'd asked for a give back, since in the previous attempt the > tests timed out. I can't reproduce this on the abel porterbox, and it > seems to me like a problem of a very busy buildd more than any bug in > the package. AFAIK the buildds are building one package at a time, and the machines where prometheus failed have 4 GB RAM. "2680815616 in use", that is 2.6 GB. You have only 3 GB address space in userspace, and I'd guess that is the limit you hit. Is this test expected to use huge amounts of memory? cu Adrian -- "Is there not promise of rain?" Ling Tan asked suddenly out of the darkness. There had been need of rain for many days. "Only a promise," Lao Er said. Pearl S. Buck - Dragon Seed ___ Pkg-go-maintainers mailing list Pkg-go-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-go-maintainers
[pkg-go] Bug#852959: Bug#852959: prometheus FTBFS on armhf: github.com/prometheus/prometheus/storage/local fails
On 28/01/17 11:58, Adrian Bunk wrote: > Source: prometheus > Version: 1.5.0+ds-1 > Severity: serious > > https://buildd.debian.org/status/fetch.php?pkg=prometheus=armhf=1.5.0%2Bds-1=1485604675=0 As it can be seen in the log: runtime: out of memory: cannot allocate 4194304-byte block (2680815616 in use) fatal error: out of memory Yesterday I'd asked for a give back, since in the previous attempt the tests timed out. I can't reproduce this on the abel porterbox, and it seems to me like a problem of a very busy buildd more than any bug in the package. -- Martín Ferrari (Tincho) ___ Pkg-go-maintainers mailing list Pkg-go-maintainers@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-go-maintainers
