On 6/27/20, Donald Wilde <dwil...@gmail.com> wrote: > On 6/26/20, Ronald Klop <ronald-li...@klop.ws> wrote: >> >> Van: Bob Bishop <r...@gid.co.uk> >> Datum: vrijdag, 26 juni 2020 17:18 >> Aan: Peter Jeremy <pe...@rulingia.com> >> CC: Donald Wilde <dwil...@gmail.com>, freebsd-stable >> <freebsd-stable@freebsd.org> >> Onderwerp: Re: swap space issues >>> >>> >>> >>> > On 26 Jun 2020, at 11:23, Peter Jeremy <pe...@rulingia.com> wrote: >>> > >>> > On 2020-Jun-25 11:30:31 -0700, Donald Wilde <dwil...@gmail.com> wrote: >>> >> Here's 'pstat -s' on the i3 (which registers as cpu HAMMER): >>> >> >>> >> Device 1K-blocks Used Avail Capacity >>> >> /dev/ada0s1b 33554432 0 33554432 0% >>> >> /dev/ada0s1d 33554432 0 33554432 0% >>> >> Total 67108864 0 67108864 0% >>> > >>> > I strongly suggest you don't have more than one swap device on >>> > spinning >>> > rust - the VM system will stripe I/O across the available devices and >>> > that will give particularly poor results when it has to seek between >>> > the >>> > partitions. >>> > Based on all recommendations on this thread (thanks, guys!), I've > rebuilt my i3 mule with exactly one 16G partition, as it has only > 'spinning rust' <haha> for a disk. My loader.conf has > kern.maxswzone=4200000 and ccache is fully active and working for both > root on tcsh and users on sh.
That appears to be successful. this is still a MBR-based system, not GPT, due to BIOS issues, but it's a 'choose my battles' situation. > > I have yet to try synth again. I'm doing buildworld/buildkernel for > 12-STABLE, but evidence so far is good. 'top -t' is actually happy, > showing 16G (grog?), so I'll try firing up synth after another hour or > so on the latest fetch of the ports tree. Top appears to have gone south again. I give up, Greg! 'pstat -s' _is_ happy. Synth is still crashing hard, same issue. As I said in freebsd-ports on my thread about 'stable postgresql11', pg11 didn't like synth's request that I create a different directory because /usr/ports/distfiles didn't exist. Doing so and adding the DISTFILES to make.conf caused all kinds of problems. I am currently (I hope) building a successful pg11-server, but I've restored /usr/ports/distfiles and before I go to bed tonight I hope to kick off another synth run. This failure happened with three builders and three tasks per builder. The symptom, multiple times, has been that llvm80 takes forever and then causes the swap failure even though pstat -s is happy AND loader conf kern.maxswzone (=4200000) is sufficient for my 16G partition. I've tried installing llvm80 as a pkg, but that didn't help. The swap stays at zero usage, and then goes almost instantaneously to OOM. The only bit of advice I'm not completely compliant with is the use of a 16G swap partition where I was advised to set max to 8G. -- Don Wilde **************************************************** * What is the Internet of Things but a system * * of systems including humans? * **************************************************** _______________________________________________ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
