From: Paolo Valente <paolo.valente@linaro.org>
To: "Srivatsa S. Bhat" <srivatsa@csail.mit.edu>
Cc: linux-fsdevel@vger.kernel.org,
linux-block <linux-block@vger.kernel.org>,
linux-ext4@vger.kernel.org, cgroups@vger.kernel.org,
linux-kernel@vger.kernel.org, axboe@kernel.dk, jack@suse.cz,
jmoyer@redhat.com, tytso@mit.edu, amakhalov@vmware.com,
anishs@vmware.com, srivatsab@vmware.com
Subject: Re: CFQ idling kills I/O performance on ext4 with blkio cgroup controller
Date: Sat, 18 May 2019 20:39:54 +0200 [thread overview]
Message-ID: <1812E450-14EF-4D5A-8F31-668499E13652@linaro.org> (raw)
In-Reply-To: <8d72fcf7-bbb4-2965-1a06-e9fc177a8938@csail.mit.edu>
[-- Attachment #1: Type: text/plain, Size: 3579 bytes --]
I've addressed these issues in my last batch of improvements for BFQ, which landed in the upcoming 5.2. If you give it a try, and still see the problem, then I'll be glad to reproduce it, and hopefully fix it for you.
Thanks,
Paolo
> Il giorno 18 mag 2019, alle ore 00:16, Srivatsa S. Bhat <srivatsa@csail.mit.edu> ha scritto:
>
>
> Hi,
>
> One of my colleagues noticed upto 10x - 30x drop in I/O throughput
> running the following command, with the CFQ I/O scheduler:
>
> dd if=/dev/zero of=/root/test.img bs=512 count=10000 oflags=dsync
>
> Throughput with CFQ: 60 KB/s
> Throughput with noop or deadline: 1.5 MB/s - 2 MB/s
>
> I spent some time looking into it and found that this is caused by the
> undesirable interaction between 4 different components:
>
> - blkio cgroup controller enabled
> - ext4 with the jbd2 kthread running in the root blkio cgroup
> - dd running on ext4, in any other blkio cgroup than that of jbd2
> - CFQ I/O scheduler with defaults for slice_idle and group_idle
>
>
> When docker is enabled, systemd creates a blkio cgroup called
> system.slice to run system services (and docker) under it, and a
> separate blkio cgroup called user.slice for user processes. So, when
> dd is invoked, it runs under user.slice.
>
> The dd command above includes the dsync flag, which performs an
> fdatasync after every write to the output file. Since dd is writing to
> a file on ext4, jbd2 will be active, committing transactions
> corresponding to those fdatasync requests from dd. (In other words, dd
> depends on jdb2, in order to make forward progress). But jdb2 being a
> kernel thread, runs in the root blkio cgroup, as opposed to dd, which
> runs under user.slice.
>
> Now, if the I/O scheduler in use for the underlying block device is
> CFQ, then its inter-queue/inter-group idling takes effect (via the
> slice_idle and group_idle parameters, both of which default to 8ms).
> Therefore, everytime CFQ switches between processing requests from dd
> vs jbd2, this 8ms idle time is injected, which slows down the overall
> throughput tremendously!
>
> To verify this theory, I tried various experiments, and in all cases,
> the 4 pre-conditions mentioned above were necessary to reproduce this
> performance drop. For example, if I used an XFS filesystem (which
> doesn't use a separate kthread like jbd2 for journaling), or if I dd'ed
> directly to a block device, I couldn't reproduce the performance
> issue. Similarly, running dd in the root blkio cgroup (where jbd2
> runs) also gets full performance; as does using the noop or deadline
> I/O schedulers; or even CFQ itself, with slice_idle and group_idle set
> to zero.
>
> These results were reproduced on a Linux VM (kernel v4.19) on ESXi,
> both with virtualized storage as well as with disk pass-through,
> backed by a rotational hard disk in both cases. The same problem was
> also seen with the BFQ I/O scheduler in kernel v5.1.
>
> Searching for any earlier discussions of this problem, I found an old
> thread on LKML that encountered this behavior [1], as well as a docker
> github issue [2] with similar symptoms (mentioned later in the
> thread).
>
> So, I'm curious to know if this is a well-understood problem and if
> anybody has any thoughts on how to fix it.
>
> Thank you very much!
>
>
> [1]. https://lkml.org/lkml/2015/11/19/359
>
> [2]. https://github.com/moby/moby/issues/21485
> https://github.com/moby/moby/issues/21485#issuecomment-222941103
>
> Regards,
> Srivatsa
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
next prev parent reply other threads:[~2019-05-18 18:40 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-05-17 22:16 CFQ idling kills I/O performance on ext4 with blkio cgroup controller Srivatsa S. Bhat
2019-05-18 18:39 ` Paolo Valente [this message]
2019-05-18 19:28 ` Theodore Ts'o
2019-05-20 9:15 ` Jan Kara
2019-05-20 10:45 ` Paolo Valente
2019-05-21 16:48 ` Theodore Ts'o
2019-05-21 18:19 ` Josef Bacik
2019-05-21 19:10 ` Theodore Ts'o
2019-05-20 10:38 ` Paolo Valente
2019-05-21 7:38 ` Andrea Righi
2019-05-18 20:50 ` Srivatsa S. Bhat
2019-05-20 10:19 ` Paolo Valente
2019-05-20 22:45 ` Srivatsa S. Bhat
2019-05-21 6:23 ` Paolo Valente
2019-05-21 7:19 ` Srivatsa S. Bhat
2019-05-21 9:10 ` Jan Kara
2019-05-21 16:31 ` Theodore Ts'o
2019-05-21 11:25 ` Paolo Valente
2019-05-21 13:20 ` Paolo Valente
2019-05-21 16:21 ` Paolo Valente
2019-05-21 17:38 ` Paolo Valente
2019-05-21 22:51 ` Srivatsa S. Bhat
2019-05-22 8:05 ` Paolo Valente
2019-05-22 9:02 ` Srivatsa S. Bhat
2019-05-22 9:12 ` Paolo Valente
2019-05-22 10:02 ` Srivatsa S. Bhat
2019-05-22 9:09 ` Paolo Valente
2019-05-22 10:01 ` Srivatsa S. Bhat
2019-05-22 10:54 ` Paolo Valente
2019-05-23 2:30 ` Srivatsa S. Bhat
2019-05-23 9:19 ` Paolo Valente
2019-05-23 17:22 ` Paolo Valente
2019-05-23 23:43 ` Srivatsa S. Bhat
2019-05-24 6:51 ` Paolo Valente
2019-05-24 7:56 ` Paolo Valente
2019-05-29 1:09 ` Srivatsa S. Bhat
2019-05-29 7:41 ` Paolo Valente
2019-05-30 8:29 ` Srivatsa S. Bhat
2019-05-30 10:45 ` Paolo Valente
2019-06-02 7:04 ` Srivatsa S. Bhat
2019-06-11 22:34 ` Srivatsa S. Bhat
2019-06-12 13:04 ` Jan Kara
2019-06-12 19:36 ` Srivatsa S. Bhat
2019-06-13 6:02 ` Greg Kroah-Hartman
2019-06-13 19:03 ` Srivatsa S. Bhat
2019-06-13 8:20 ` Jan Kara
2019-06-13 19:05 ` Srivatsa S. Bhat
2019-06-13 8:37 ` Jens Axboe
2019-06-13 5:46 ` Paolo Valente
2019-06-13 19:13 ` Srivatsa S. Bhat
2019-05-23 23:32 ` Srivatsa S. Bhat
2019-05-30 8:38 ` Srivatsa S. Bhat
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1812E450-14EF-4D5A-8F31-668499E13652@linaro.org \
--to=paolo.valente@linaro.org \
--cc=amakhalov@vmware.com \
--cc=anishs@vmware.com \
--cc=axboe@kernel.dk \
--cc=cgroups@vger.kernel.org \
--cc=jack@suse.cz \
--cc=jmoyer@redhat.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=srivatsa@csail.mit.edu \
--cc=srivatsab@vmware.com \
--cc=tytso@mit.edu \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).