On 2019-03-03 00:15:13 +0100, Guillem Jover wrote:
> Did this start happening due to the new kernel version?
In March 2018, this was already slow, but much less:
[...]
2018-03-08 00:55:55 status unpacked texlive-fonts-extra:all 2017.20180225-1
2018-03-08 00:55:55 status half-installed texlive-fonts-extra:all
2017.20180225-1
2018-03-08 01:02:55 status half-installed texlive-fonts-extra:all
2017.20180225-1
2018-03-08 01:02:56 status unpacked texlive-fonts-extra:all 2017.20180305-2
2018-03-08 01:02:56 status unpacked texlive-fonts-extra:all 2017.20180305-2
[...]
I don't have older information.
> What filesystem are you using (say btrfs or similar?),
mount information:
/dev/mapper/zira--vg-root on / type ext4 (rw,relatime,errors=remount-ro)
> etc, can you reproduce on another system? I certainly can't. :)
No, on two of my other machines, it takes around 1 minute
(both for old and new versions). I assume that this is what
is expected.
> I guess attaching strace to the running dpkg might also give some more
> info on what's going on, or where it's stalling. Also if you can
> simply reproduce via «dpkg -i», then enabling all dpkg debug flags
> might also help.
With
dpkg -i /var/cache/apt/archives/texlive-fonts-extra_2018.20190227-1_all.deb
GKrellM shows full disk activity for a bit less than 2 minutes, then
no activity, but dpkg is still running. Attaching strace against it
during this time shows lots of rename's:
[...]
rename("/usr/share/texlive/texmf-dist/fonts/tfm/arkandis/berenisadf/ybdr2cijw8y.tfm.dpkg-new",
"/usr/share/texlive/texmf-dist/fonts/tfm/arkandis/berenisadf/ybdr2cijw8y.tfm")
= 0
openat(AT_FDCWD,
"/usr/share/texlive/texmf-dist/fonts/tfm/arkandis/berenisadf/ybdr2ciw8y.tfm.dpkg-new",
O_WRONLY) = 11
fsync(11) = 0
close(11) = 0
rename("/usr/share/texlive/texmf-dist/fonts/tfm/arkandis/berenisadf/ybdr2ciw8y.tfm.dpkg-new",
"/usr/share/texlive/texmf-dist/fonts/tfm/arkandis/berenisadf/ybdr2ciw8y.tfm")
= 0
openat(AT_FDCWD,
"/usr/share/texlive/texmf-dist/fonts/tfm/arkandis/berenisadf/ybdr2cj8c.tfm.dpkg-new",
O_WRONLY) = 11
fsync(11) = 0
close(11) = 0
rename("/usr/share/texlive/texmf-dist/fonts/tfm/arkandis/berenisadf/ybdr2cj8c.tfm.dpkg-new",
"/usr/share/texlive/texmf-dist/fonts/tfm/arkandis/berenisadf/ybdr2cj8c.tfm") = 0
openat(AT_FDCWD,
"/usr/share/texlive/texmf-dist/fonts/tfm/arkandis/berenisadf/ybdr2cj8t.tfm.dpkg-new",
O_WRONLY) = 11
fsync(11) = 0
close(11) = 0
[...]
> > -- System Information:
> > Debian Release: buster/sid
> > APT prefers unstable-debug
> > APT policy: (500, 'unstable-debug'), (500, 'stable-updates'), (500,
> > 'unstable'), (500, 'testing'), (500, 'stable'), (1, 'experimental')
> > Architecture: amd64 (x86_64)
> >
> > Kernel: Linux 4.19.0-3-amd64 (SMP w/8 CPU cores)
> > Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE,
> > TAINT_UNSIGNED_MODULE
> > Locale: LANG=POSIX, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE=POSIX
> > (charmap=UTF-8)
> > Shell: /bin/sh linked to /bin/dash
> > Init: systemd (via /run/systemd/system)
> > LSM: AppArmor: enabled
>
> I see the kernel is tainted. Perhaps try to unload those and see? :)
That's the NVIDIA drivers (because nouveau is too broken).
--
Vincent Lefèvre <vinc...@vinc17.net> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
