[systemd-devel] Wtrlt: Re: Antw: [EXT] Re: What causes "systemd-journald[3256]: Missed 127 kernel messages"

[Ulrich Windl]

(forgot the list)
>>> Ulrich Windl <Ulrich.Windl at rz.uni-regensburg.de> schrieb am 01.06.2021 um
14:57
in Nachricht <60B64B0C.ED38.00A1.0 at rz.uni-regensburg.de>:
>>>> Lennart Poettering <lennart at poettering.net> schrieb am 01.06.2021 um
14:43 in
> Nachricht <YLYrch+r3wc+QqO5 at gardel-login>:
> > On Di, 01.06.21 14:33, Ulrich Windl (Ulrich.Windl at rz.uni-regensburg.de) 
> > wrote:
> > 
> >> >>> Lennart Poettering <lennart at poettering.net> schrieb am 01.06.2021 um
13:39
> >> in
> >> Nachricht <YLYcdMZ+MgmcPdmG at gardel-login>:
> >> > On Di, 01.06.21 12:42, Ulrich Windl
(Ulrich.Windl at rz.uni‑regensburg.de)
> >> wrote:
> >> >
> >> >> Jun 01 12:33:10 h18 systemd‑journald[3256]: Missed 195 kernel
messages
> >> >>
> >> >> A few questions:
> >> >> 1) What causes this?
> >> >
> >> > Dunno. Something is massively flooding the kernel log buffer. Probably
> >> > some borked driver or so. "dmesg" might tell you what.
> >>
> >> I had meant the dropping of messages, not the creation of such. It seems

> > it's
> >> intentional.
> > 
> > Ahumm. It does not. Generating such high frequency log messages is a
> > bug. Please report to your kernel maintainers.
> 
> OK, maybe read again the beginning of my message: I had >500 processes in 
> "D" state when doing "echo w >/proc/sysrq-trigger".
> 
> > 
> >> > You could also enlarge the kernel log buffer, see log_buf_mem= kernel
> >> > cmdline switch.
> >>
> >> Confused: So is it the kernel dropping/loosing messages, or is it
journald?
> > 
> > The kernel is generating them faster than userspace can keep up with
them.
> 
> That's bad.
> 
> > 
> > Lennart
> > 
> > --
> > Lennart Poettering, Berlin
> 
> 
> 
>