From: Steven Haigh <netwiz@crc.id.au>
To: linux-kernel <linux-kernel@vger.kernel.org>
Subject: Fedora 26 / Kernel 4.11.10 - System becomes unresponsive when swapping
Date: Tue, 18 Jul 2017 09:10:31 +1000
Message-ID: <1540730.ipWl7zM1Wl@wopr.lan.crc.id.au>
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="nextPart3095397.ilJzbD7srN"; micalg="pgp-sha256"; protocol="application/pgp-signature"
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 3321
Lines: 90

--nextPart3095397.ilJzbD7srN
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="UTF-8"

Hi all,

=46irstly, please CC me directly in replies due to not being subscribed.

After a fix to glibc that fixed Unity 3D based games (Fedora ref: https://
bugzilla.redhat.com/show_bug.cgi?id=3D1440287), I have noticed that when I =
play=20
Cities: Skylines that the system becomes unresponsive when digging into swa=
p.

I have 10Gb of RAM in this system and run Fedora 26. If I launch Cities:=20
Skylines with no swap space, things run well performance wise until I get a=
n=20
OOM - and it all dies - which is expected.

When I turn on swap to /dev/sda2 which resides on an SSD, I get complete=20
system freezes while swap is being accessed.

The first swap was after loading a saved game, then launching kmail in the=
=20
background. This caused ~500Mb to be swapped to /dev/sda2 on an SSD. The=20
system froze for about 8 minutes - barely being able to move the mouse. The=
=20
HDD LED was on constantly during the entire time.

To hopefully rule out the above glibc issue, I started the game via jemallo=
c -=20
but experienced even more severe freezes while swapping. I gave up waiting=
=20
after 13 minutes of non-responsiveness - not even being able to move the mo=
use=20
properly.

During these hangs, I could typed into a Konsole window, and some of the=20
typing took 3+ minutes to display on the screen (yay for buffers?).

I have tested this with both the default vm.swappiness values, as well as t=
he=20
following:
vm.swappiness =3D 1
vm.min_free_kbytes =3D 32768
vm.vfs_cache_pressure =3D 60

I noticed that when I do eventually get screen updates, all 8 cpus (4 cores=
 /=20
2 threads) show 100% CPU usage - and kswapd is right up there in the proces=
s=20
list for CPU usage. Sadly I haven't been able to capture this information=20
fully yet due to said unresponsiveness.

This seems to be a relatively new problem that I did not encounter during t=
he=20
=46edora 26 beta - but do now.

Does anyone have any suggestions on where to start tracking this type of is=
sue=20
down?

=2D-=20
Steven Haigh

=F0=9F=93=A7 netwiz@crc.id.au       =F0=9F=92=BB http://www.crc.id.au
=F0=9F=93=9E +61 (3) 9001 6090    =F0=9F=93=B1 0412 935 897
--nextPart3095397.ilJzbD7srN
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: This is a digitally signed message part.
Content-Transfer-Encoding: 7Bit

-----BEGIN PGP SIGNATURE-----
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=N6E6
-----END PGP SIGNATURE-----

--nextPart3095397.ilJzbD7srN--