Created attachment 693696 [details] build.log and emerge --info build.log and emerge --info attached * ERROR: dev-lang/rust-1.50.0::gentoo failed (compile phase): * (no error message) * * Call stack: * ebuild.sh, line 125: Called src_compile * environment, line 3508: Called die * The specific snippet of code: * env $(cat "${S}"/config.env) RUST_BACKTRACE=1 "${EPYTHON}" ./x.py dist -vv --config="${S}"/config.toml -j$(makeopts_jobs) || die ) * * If you need support, post the output of `emerge --info '=dev-lang/rust-1.50.0::gentoo'`, * the complete build log and the output of `emerge -pqv '=dev-lang/rust-1.50.0::gentoo'`. * The complete build log is located at '/var/tmp/portage/dev-lang/rust-1.50.0/temp/build.log'. * The ebuild environment file is located at '/var/tmp/portage/dev-lang/rust-1.50.0/temp/environment'. * Working directory: '/var/tmp/portage/dev-lang/rust-1.50.0/work/rustc-1.50.0-src' * S: '/var/tmp/portage/dev-lang/rust-1.50.0/work/rustc-1.50.0-src'
rustc exited with signal: 9 That implies something killed the process, most likely the kernel's OOM (Out-of-Memory) killer, so you likely ran out of RAM. Would be mentioned near the end of `dmesg`'s output. I suggest to use lower MAKEOPTS -jN value (like halved to -j6) for this package, rustc can sometime use 2-3.5GB per thread and it adds up. Also ensure not using tmpfs for PORTAGE_TMPDIR). (on side-note, preferably don't use tar for logs)
agreed. try to get 2GB per thread for C++ and rust code. so with 18GB it should be no more than -j9 and much much lower than that if you use TMPFS build dir. if you don't use system-bootstrap and system-llvm, tmpfs will use approx 9GB to unpack and build it. that leaves you with 9g to work with. assuming you need at least 1G for runtime, I would not set more than -j4 to build rust with 18G and tmpfs.
Created attachment 693792 [details] new build.log
(In reply to Ionen Wolkens from comment #1) > rustc exited with signal: 9 > > That implies something killed the process, most likely the kernel's OOM > (Out-of-Memory) killer, so you likely ran out of RAM. Would be mentioned > near the end of `dmesg`'s output. > > I suggest to use lower MAKEOPTS -jN value (like halved to -j6) for this > package, rustc can sometime use 2-3.5GB per thread and it adds up. Also > ensure not using tmpfs for PORTAGE_TMPDIR). > > (on side-note, preferably don't use tar for logs) Ah okay, my bad I thought I had plenty free memory when building but maybe some was reserved. I just did a clean reboot and re-emerged rust and it had 17.5GB of memory to build with and it still failed. I know you said to reduce my jobs but I wanted to be sure it was the memory because I've never had that problem in the past when I built it. I've attached the new build.log, I checked it and didn't see any oom errors, but it could be that I just don't know what they look like (I was just skimming through/searching in vim).
(In reply to Georgy Yakovlev from comment #2) > agreed. > try to get 2GB per thread for C++ and rust code. > > so with 18GB it should be no more than -j9 and much much lower than that if > you use TMPFS build dir. > > > if you don't use system-bootstrap and system-llvm, tmpfs will use approx 9GB > to unpack and build it. > that leaves you with 9g to work with. assuming you need at least 1G for > runtime, I would not set more than -j4 to build rust with 18G and tmpfs. I didn't account for my tmpfs... I'll reduce my jobs and try again. Sorry to be a bother. I really appreciate the feedback though!
It built when I set it to -j3.
Glad to hear it works.
