[Bug 277389] Reproduceable low memory freeze on 14.0-RELEASE-p5

From: <bugzilla-noreply_at_freebsd.org>
Date: Sun, 03 Nov 2024 00:45:55 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=277389

--- Comment #42 from Mark Millard <marklmi26-fbsd@yahoo.com> ---
I just checked a more recent PkgBase kernel & world based
system on the 32 GiByte Windows Dev Kit 2023 without
swap space being enabled. USB3 UFS boot media with separate
USB3 ZFS media imported later. (Each: U.2 adapter to large
Optane media.)

The read activity quickly got OOM activity. (The
iozone.DUMMY.* files existed from a prior boot and were
reused.) My personal variant of top showed:

Mem: 132568Ki Active, 250168Ki Inact, 30802Mi Wired, 248328Ki Buf, 116268Ki
Free, 132568Ki MaxObsActive, 30802Mi MaxObsWired, 30932Mi MaxObs(Act+Wir+Lndry)
ARC: 28008Mi Total, 1963Mi MFU, 25813Mi MRU, 79723Ki Header, 158558Ki Other
     26405Mi Compressed, 27557Mi Uncompressed, 1.04:1 Ratio

as of when it was killed. Note that "Max(imum)Obs(erved)..."
figures ones that I added and I changed the number/units
output style: more significant digits and explicitly powers
of 1024 based.

Note the "30802Mi Wired" and "ARC: 28008Mi Total" and
"25813Mi MRU" and "26405Mi Compressed": the figures are
not staying in reasonable bounds for the RAM available.

This was from:

# iozone -w -i 1 -l 512 -r 4k -s 1g

The kills left me without any normal access, including via
the (plain text) console and via ssh. But I could break
into ddb via the console keyboard.

. . .
Nov  2 16:08:16 aarch64-main-pbase kernel: pid 4639 (top), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:08:33 aarch64-main-pbase kernel: pid 4640 (sshd), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:08:46 aarch64-main-pbase kernel: pid 4630 (sshd), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:08:59 aarch64-main-pbase kernel: pid 4353 (ntpd), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:09:12 aarch64-main-pbase kernel: pid 4515 (sh), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:09:24 aarch64-main-pbase kernel: pid 4085 (devd), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:09:35 aarch64-main-pbase kernel: pid 4507 (login), jid 0, uid 0, was
killed: a thread waited too long to allocate a page
Nov  2 16:09:37 aarch64-main-pbase kernel: pid 4443 (mountd), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:09:56 aarch64-main-pbase kernel: pid 4316 (rpcbind), jid 0, uid 0,
was killed: failed to reclaim memory
Nov  2 16:10:09 aarch64-main-pbase kernel: pid 127 (adjkerntz), jid 0, uid 0,
was killed: failed to reclaim memory
Nov  2 16:10:21 aarch64-main-pbase kernel: pid 3887 (dhclient), jid 0, uid 65,
was killed: failed to reclaim memory
Nov  2 16:10:31 aarch64-main-pbase dhclient[3778]: connection closed
Nov  2 16:10:34 aarch64-main-pbase kernel: pid 3781 (dhclient), jid 0, uid 0,
was killed: failed to reclaim memory
Nov  2 16:10:41 aarch64-main-pbase dhclient[3778]: exiting.
Nov  2 16:10:47 aarch64-main-pbase kernel: pid 3778 (dhclient), jid 0, uid 0,
was killed: failed to reclaim memory
Nov  2 16:10:59 aarch64-main-pbase kernel: pid 4513 (getty), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:11:12 aarch64-main-pbase kernel: pid 4512 (getty), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:11:25 aarch64-main-pbase kernel: pid 4510 (getty), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:11:36 aarch64-main-pbase kernel: pid 4508 (getty), jid 0, uid 0, was
killed: a thread waited too long to allocate a page
Nov  2 16:11:38 aarch64-main-pbase kernel: pid 4514 (getty), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:11:50 aarch64-main-pbase kernel: pid 4511 (getty), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:12:01 aarch64-main-pbase kernel: pid 4509 (getty), jid 0, uid 0, was
killed: a thread waited too long to allocate a page
Nov  2 16:12:03 aarch64-main-pbase kernel: pid 4037 (moused), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:12:16 aarch64-main-pbase kernel: pid 4445 (nfsd), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:12:28 aarch64-main-pbase kernel: pid 4446 (nfsd), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:12:41 aarch64-main-pbase kernel: pid 5173 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:12:54 aarch64-main-pbase kernel: pid 5172 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:13:03 aarch64-main-pbase kernel: ugen0.6: <PixArt USB Optical Mouse>
at usbus0 (disconnected)
Nov  2 16:13:03 aarch64-main-pbase kernel: ums0: at uhub3, port 2, addr 5
(disconnected)
Nov  2 16:13:03 aarch64-main-pbase kernel: ums0: detached
Nov  2 16:13:05 aarch64-main-pbase kernel: ugen0.6: <PixArt USB Optical Mouse>
at usbus0
Nov  2 16:13:05 aarch64-main-pbase kernel: ums0 on uhub3
Nov  2 16:13:05 aarch64-main-pbase kernel: ums0: <PixArt USB Optical Mouse,
class 0/0, rev 1.10/1.00, addr 5> on usbus0
Nov  2 16:13:05 aarch64-main-pbase kernel: ums0: 3 buttons and [XYZ]
coordinates ID=0
Nov  2 16:13:06 aarch64-main-pbase kernel: pid 5171 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:13:19 aarch64-main-pbase kernel: pid 5170 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:13:32 aarch64-main-pbase kernel: pid 5169 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:13:44 aarch64-main-pbase kernel: pid 5168 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:13:57 aarch64-main-pbase kernel: pid 5167 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
. . .
Nov  2 16:43:09 aarch64-main-pbase kernel: pid 4986 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:43:22 aarch64-main-pbase kernel: pid 4985 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:43:34 aarch64-main-pbase kernel: pid 4984 (iozone), jid 0, uid 0, was
killed: a thread waited too long to allocate a page
Nov  2 16:43:35 aarch64-main-pbase kernel: pid 4983 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:43:47 aarch64-main-pbase kernel: pid 4982 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
Nov  2 16:43:58 aarch64-main-pbase kernel: pid 4981 (iozone), jid 0, uid 0, was
killed: a thread waited too long to allocate a page
Nov  2 16:43:58 aarch64-main-pbase kernel: ugen0.6: <PixArt USB Optical Mouse>
at usbus0 (disconnected)
Nov  2 16:43:58 aarch64-main-pbase kernel: ums0: at uhub3, port 2, addr 5
(disconnected)
Nov  2 16:43:58 aarch64-main-pbase kernel: ums0: detached
Nov  2 16:43:59 aarch64-main-pbase kernel: ugen0.6: <PixArt USB Optical Mouse>
at usbus0
Nov  2 16:43:59 aarch64-main-pbase kernel: ums0 on uhub3
Nov  2 16:43:59 aarch64-main-pbase kernel: ums0: <PixArt USB Optical Mouse,
class 0/0, rev 1.10/1.00, addr 5> on usbus0
Nov  2 16:43:59 aarch64-main-pbase kernel: ums0: 3 buttons and [XYZ]
coordinates ID=0
Nov  2 16:44:00 aarch64-main-pbase kernel: pid 4980 (iozone), jid 0, uid 0, was
killed: failed to reclaim memory
. . .

Definitely a denial of service context.

For reference:

# uname -apKU
FreeBSD aarch64-main-pbase 15.0-CURRENT FreeBSD 15.0-CURRENT
main-n273174-8b2e7da70855 GENERIC-NODEBUG arm64 aarch64 1500026 1500026

So this is after the fix to the system accounting of laundry to
span both swappable and unswappable for the sysctl that top
is based on for such.

-- 
You are receiving this mail because:
You are on the CC list for the bug.