Buildbot failure in Wildebeest Builder on whole buildset
The Buildbot has detected a new failure on builder gccrust-debian-arm64 while building gccrust. Full details are available at: https://builder.wildebeest.org/buildbot/#builders/58/builds/1648 Buildbot URL: https://builder.wildebeest.org/buildbot/ Worker for this Build: debian-arm64 Build Reason: Blamelist: Philip Herron BUILD FAILED: failed compile (failure) Sincerely, -The Buildbot -- Gcc-rust mailing list Gcc-rust@gcc.gnu.org https://gcc.gnu.org/mailman/listinfo/gcc-rust
Re: Buildbot failure in Wildebeest Builder on whole buildset
build...@builder.wildebeest.org writes: > Build Reason: > Blamelist: Philip Herron g++ -fno-PIE -c -DIN_GCC_FRONTEND -g -O2 -DIN_GCC -fno-exceptions -fno-rtti -fasynchronous-unwind-tables -W -Wall -Wno-narrowing -Wwrite-strings -Wcast-qual -Wmissing-format-attribute -Woverloaded-virtual -pedantic -Wno-long-long -Wno-variadic-macros -Wno-overlength-strings -fno-common -DHAVE_CONFIG_H -I. -Irust -I../../gccrs/gcc -I../../gccrs/gcc/rust -I../../gccrs/gcc/../include -I../../gccrs/gcc/../libcpp/include -I../../gccrs/gcc/../libcody -I../../gccrs/gcc/../libdecnumber -I../../gccrs/gcc/../libdecnumber/dpd -I../libdecnumber -I../../gccrs/gcc/../libbacktrace -o rust/rust-hir-full-test.o -MT rust/rust-hir-full-test.o -MMD -MP -MF rust/.deps/rust-hir-full-test.TPo -std=c++11 -Wno-unused-parameter -Werror=overloaded-virtual -I ../../gccrs/gcc/rust -I ../../gccrs/gcc/rust/lex -I ../../gccrs/gcc/rust/parse -I ../../gccrs/gcc/rust/ast -I ../../gccrs/gcc/rust/analysis -I ../../gccrs/gcc/rust/backend -I ../../gccrs/gcc/rust/expand -I ../../gccrs/gcc/rust/hir/tree -I ../ ../gccrs/gcc/rust/hir -I ../../gccrs/gcc/rust/resolve -I ../../gccrs/gcc/rust/util -I ../../gccrs/gcc/rust/typecheck -I ../../gccrs/gcc/rust/lint ../../gccrs/gcc/rust/hir/tree/rust-hir-full-test.cc command timed out: 1200 seconds without output running [b'make', b'-j4'], attempting to kill process killed by signal 9 program finished with exit code -1 elapsedTime=1425.203557 Looks suspicious. Is it possible that the VM got stuck ? Marc -- Gcc-rust mailing list Gcc-rust@gcc.gnu.org https://gcc.gnu.org/mailman/listinfo/gcc-rust
Re: Buildbot failure in Wildebeest Builder on whole buildset
Hi Marc, On Thu, Feb 17, 2022 at 08:46:01PM +0100, Marc via Gcc-rust wrote: > build...@builder.wildebeest.org writes: > > > Build Reason: > > Blamelist: Philip Herron > > g++ -fno-PIE -c -DIN_GCC_FRONTEND -g -O2 -DIN_GCC -fno-exceptions > -fno-rtti -fasynchronous-unwind-tables -W -Wall -Wno-narrowing > -Wwrite-strings -Wcast-qual -Wmissing-format-attribute -Woverloaded-virtual > -pedantic -Wno-long-long -Wno-variadic-macros -Wno-overlength-strings > -fno-common -DHAVE_CONFIG_H -I. -Irust -I../../gccrs/gcc > -I../../gccrs/gcc/rust -I../../gccrs/gcc/../include > -I../../gccrs/gcc/../libcpp/include -I../../gccrs/gcc/../libcody > -I../../gccrs/gcc/../libdecnumber -I../../gccrs/gcc/../libdecnumber/dpd > -I../libdecnumber -I../../gccrs/gcc/../libbacktrace -o > rust/rust-hir-full-test.o -MT rust/rust-hir-full-test.o -MMD -MP -MF > rust/.deps/rust-hir-full-test.TPo -std=c++11 -Wno-unused-parameter > -Werror=overloaded-virtual -I ../../gccrs/gcc/rust -I > ../../gccrs/gcc/rust/lex -I ../../gccrs/gcc/rust/parse -I > ../../gccrs/gcc/rust/ast -I ../../gccrs/gcc/rust/analysis -I > ../../gccrs/gcc/rust/backend -I ../../gccrs/gcc/rust/expand -I > ../../gccrs/gcc/rust/hir/tree -I . ./ > ../gccrs/gcc/rust/hir -I ../../gccrs/gcc/rust/resolve -I > ../../gccrs/gcc/rust/util -I ../../gccrs/gcc/rust/typecheck -I > ../../gccrs/gcc/rust/lint ../../gccrs/gcc/rust/hir/tree/rust-hir-full-test.cc > command timed out: 1200 seconds without output running [b'make', b'-j4'], > attempting to kill > process killed by signal 9 > program finished with exit code -1 > elapsedTime=1425.203557 > > Looks suspicious. Is it possible that the VM got stuck ? It did indeed. The machine only has 4GB of memory and 2GB swap. With make -j4 there are 4 processes each taking ~1GB and it is just swapping: 926 root 20 0 0 0 0 R 97.6 0.0 135:41.18 kswapd0 15794 buildbot 20 0 1539436 928056 0 D 2.4 24.5 1:41.92 cc1plus 15817 buildbot 20 0 781108 674280328 D 2.4 17.8 1:08.31 cc1plus 15792 buildbot 20 0 1250264 763912384 D 2.1 20.2 1:40.62 cc1plus 15800 buildbot 20 0 1427104 784820132 D 2.1 20.7 1:41.67 cc1plus I'll reconfigure with make -j3 to see if that helps. I don't know why we are suddenly seeing this, did the code get that much bigger recently? Cheers, Mark -- Gcc-rust mailing list Gcc-rust@gcc.gnu.org https://gcc.gnu.org/mailman/listinfo/gcc-rust
Re: Buildbot failure in Wildebeest Builder on whole buildset
Hi, On Thu, Feb 17, 2022 at 10:05:29PM +0100, Mark Wielaard wrote: > On Thu, Feb 17, 2022 at 08:46:01PM +0100, Marc via Gcc-rust wrote: > > ../gccrs/gcc/rust/hir -I ../../gccrs/gcc/rust/resolve -I > > ../../gccrs/gcc/rust/util -I ../../gccrs/gcc/rust/typecheck -I > > ../../gccrs/gcc/rust/lint > > ../../gccrs/gcc/rust/hir/tree/rust-hir-full-test.cc > > command timed out: 1200 seconds without output running [b'make', b'-j4'], > > attempting to kill > > process killed by signal 9 > > program finished with exit code -1 > > elapsedTime=1425.203557 > > > > Looks suspicious. Is it possible that the VM got stuck ? > > It did indeed. The machine only has 4GB of memory and 2GB swap. With > make -j4 there are 4 processes each taking ~1GB and it is just > swapping: > > 926 root 20 0 0 0 0 R 97.6 0.0 135:41.18 kswapd0 > 15794 buildbot 20 0 1539436 928056 0 D 2.4 24.5 1:41.92 cc1plus > 15817 buildbot 20 0 781108 674280328 D 2.4 17.8 1:08.31 cc1plus > 15792 buildbot 20 0 1250264 763912384 D 2.1 20.2 1:40.62 cc1plus > 15800 buildbot 20 0 1427104 784820132 D 2.1 20.7 1:41.67 cc1plus > > I'll reconfigure with make -j3 to see if that helps. It didn't. With -j3 the oom-killer kicks in (dunno why not with -j4): [1643873.684998] Out of memory: Killed process 16625 (cc1plus) total-vm:1764740kB, anon-rss:1040744kB, file-rss:0kB, shmem-rss:0kB, UID:998 pgtables:3464kB oom_score_adj:0 [1643873.992761] oom_reaper: reaped process 16625 (cc1plus), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB So now using make -j2 which does succeed. https://builder.wildebeest.org/buildbot/#/builders/58/builds/1655 Very odd. But apparently for some reason 4GB isn't enough anymore to use -j3 or -j4. Which is unfortunate because this board doesn't take more than 4GB memory. Cheers, Mark -- Gcc-rust mailing list Gcc-rust@gcc.gnu.org https://gcc.gnu.org/mailman/listinfo/gcc-rust
