i didn't get anywhere close to finding the real source, but here is my observation: * system stops to be responsive when heavy writes are perfomed (like when unpacking files during emerge of sci-libs/vtk) * system stops to be responsive during compilation of sci-libs/vtk even with MAKEOPTS=-j1 for some packages i did not encounter this behavior, but with sci-libs/vtk it's a sure bet. the underlying system is btrfs raid1 on nvme disks. with gentoo-sources-6.9.9 there is no such issue.
it just happened to me with 6.10.1 too, with two compilations running in parallel (freecad and tomcat), the system became unresponsive.
Can you do a git bisect between the last working kernel and the first non-working kernel.
(In reply to Mike Pagano from comment #2) > Can you do a git bisect between the last working kernel and the first > non-working kernel. i'll try to find some time but it's hard these days. anyway, the 6.9.x branch seems to not be affected, now on 6.9.11 and no issues so far. what would be the best way to bisect it? should i just use the vanilla kernel?
(In reply to Miroslav Šulc from comment #3) > (In reply to Mike Pagano from comment #2) > > Can you do a git bisect between the last working kernel and the first > > non-working kernel. > > i'll try to find some time but it's hard these days. anyway, the 6.9.x > branch seems to not be affected, now on 6.9.11 and no issues so far. > > what would be the best way to bisect it? should i just use the vanilla > kernel? Stay on 6.9.11, and let me know if it happens again. Might have been fixed already.
6.9.11 seems to be fine, i just tried emerging vtk for quite some time and not a sign of any issues, so it seems 6.9.11 is not affected, as well as 6.9.9 was not affected.
just tried 6.10.2 - still affected
(In reply to Miroslav Šulc from comment #6) > just tried 6.10.2 - still affected Report it to the btrfs mailing list, that's the only way to get meaningful help.
(In reply to Holger Hoffstätte from comment #7) > (In reply to Miroslav Šulc from comment #6) > > just tried 6.10.2 - still affected > > Report it to the btrfs mailing list, that's the only way to get meaningful > help. i'm not really sure it's btrfs related. and it's only my guess that heavy writing causes the issue, as i encouter some freezes during unpacking during emerge and also during compilation. that's the part where the system becomes completely unresponsive. but it's usually hard to tell as i can do nothing when the system stops to respond...
i've been running 6.10.3 for several hours now with some heavy compilations and so far no issue.
the issue seems not to be fully resolved yet in 6.10.3 but it's already much better.
https://lore.kernel.org/regressions/CAL3q7H4iYRsjG9BvRYh_aB6UN-QFuTCqJdiq6hV_Xh7+U7qJ5A@mail.gmail.com/T/#t / https://bugzilla.kernel.org/show_bug.cgi?id=219121 maybe
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=ae1e766f623f7a2a889a0b09eb076dd9a60efbe9
(In reply to Mike Pagano from comment #12) > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/ > ?id=ae1e766f623f7a2a889a0b09eb076dd9a60efbe9 This commit, not yet released in any kernels, is purported to address system freezing/locking up during high IO usage. Test at your own risk, of course.
This patch is is 6.10.7. Please test with that version and let us know if this problem still occurs.
(In reply to Mike Pagano from comment #14) > This patch is is 6.10.7. Please test with that version and let us know if > this problem still occurs. i did some heavy compilations and so far no issues.
Ok, good. Sounds like this is addressed. Please re-open if that turns out to be not the case.
yeah, it seems to be resolved, no issues since 6.10.7
