Date: Sat, 11 Apr 2020 18:41:37 +0000 (UTC) From: Chris Ross <cross+freebsd@distal.com> To: Robert Clausecker <fuz@fuz.su> Cc: freebsd-fs <freebsd-fs@freebsd.org>, Chris Ross <cross+freebsd@distal.com> Subject: Re: ZFS server has gone crazy slow Message-ID: <A1B20EE1-8F51-4A87-9713-31D60B2CA1B4@distal.com> In-Reply-To: <6190573D-BCA7-44F9-86BD-0DCBB1F69D1D@distal.com> References: <2182C27C-A5D3-41BF-9CE9-7C6883E43074@distal.com> <20200411174831.GA54397@fuz.su> <6190573D-BCA7-44F9-86BD-0DCBB1F69D1D@distal.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> On Apr 11, 2020, at 14:02, Chris Ross <cross+freebsd@distal.com> wrote:
> I _think_ this is a filesystem problem. [=E2=80=A6]
Okay, this is pretty certain now. I=E2=80=99m trying to disable some daemo=
ns (squid specifically), and to move a large log file out of the way, I did=
:
% mv access.log access.log.old && touch access.log
...as the squid user. For 10 minutes or more, that mv was in disk wait:
squid 2365 0.0 0.0 4240 1892 3 D+ 14:22 0:00.02 mv access=
.log access.log.old
I didn=E2=80=99t truss that mv to know why, but that should never happen. =
A rename in the same directory is about the most simple action to perform=
=E2=80=A6
I=E2=80=99m going to reboot now that I=E2=80=99ve disabled squid and moved =
it=E2=80=99s logs (it did eventually complete, somewhere in the 11-14 minut=
e range). I asked the daemon to stop 30+ minutes ago, it still hasn=E2=80=
=99t exited=E2=80=A6. I doubt that will fix the issues I=E2=80=99m having, =
but it=E2=80=99s a start.
- Chris
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?A1B20EE1-8F51-4A87-9713-31D60B2CA1B4>