From: Vlastimil Babka <vbabka@suse.cz>
To: Mike Galbraith <efault@gmx.de>, linux-mm <linux-mm@kvack.org>
Cc: Hugh Dickins <hughd@google.com>, Ivan Kalvachev <ikalvachev@gmail.com>
Subject: Re: KVM OS update induced thrashing
Date: Fri, 1 Jun 2018 10:09:08 +0200 [thread overview]
Message-ID: <4505a137-b4e1-4c2a-67d0-3d526fac0503@suse.cz> (raw)
In-Reply-To: <1527674684.5297.90.camel@gmx.de>
On 05/30/2018 12:04 PM, Mike Galbraith wrote:
> Greetings mm wizards,
>
> Yesterday, while running master-rc7, I started updating one of my VM
> images from openSUSE Leap 42.3 -> Leap 15, figuring that could chug
> along while I do some work. That turned out to be a bad idea, box
> thrashed so badly it became unusable, which seems worthy of mention.
>
> Box is a garden variety i4790 box with 16G ram, half of which was
> donated to the VM. The VM image is 64GB, it's an OS plus all critical
> data mirror of my box (that fits on a USB stick). All was well until
> qemu RSS grew to the full 8G, at which time mm land went mad, swapping
> and thrashing horribly while 7G of ram remained free, see vmstat
> snippet below.
Could it be the same issue that Hugh [1] and Ivan [2] were patching
recently?
Vlastimil
[1] https://marc.info/?l=linux-mm&m=152773700811689&w=2
[2] https://marc.info/?l=linux-mm&m=152779529010674
> It's NOT fully repeatable. I repeated the update using a 4.14 kernel,
> which behaved fine. Switching to 4.16, box thrashed, but not as badly
> as master had, nor did it keep as much ram free while doing so (~3G).
If there's still a problem with 4.16 then perhaps not, or there's also
another issue in addition to the 4.17 one...
> Going back to master, mm land remained calm, I could use the box to do
> merges and whatnot (albeit being careful to take it easy on spinning
> rust) just fine while the VM did it's thing. Phase-of-moon behavior.
>
> This master vmstat snippet is with only the VM and a mostly idle GUI,
> as trying to do anything resembling work was a waste of time.
> procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
> r b swpd free buff cache si so bi bo in cs us sy id wa st
> 1 1 1740288 6896088 6828 128340 21 29 350 367 122 103 7 2 83 8 0
> 0 2 1740032 6902320 6828 128636 80 0 1392 18332 2429 8460 8 1 84 7 0
> 1 0 1740032 6898392 6828 128628 28 0 748 1640 2714 5838 6 2 87 6 0
> 1 0 1739776 6897864 6828 128668 164 0 1920 76 2558 7639 3 1 90 6 0
> 1 0 1739776 6895372 6828 129156 116 0 1896 60 2104 6456 3 1 91 5 0
> 0 0 1739776 6895116 6836 129192 196 0 700 168 2945 6677 7 1 89 4 0
> 0 1 1739776 6901092 6844 129292 40 0 1280 1572 3687 9739 5 3 86 6 0
> 1 0 1739776 6898884 6844 128236 12 64 276 1872 4596 18786 8 3 86 3 0
> 0 2 1747200 6910216 6760 127608 8 7380 288 7748 6725 38107 9 4 84 3 0
> 0 0 1756932 6918516 6724 126016 4 10064 172 10420 4711 10038 6 2 88 3 0
> 0 0 1773824 6936980 5728 125060 16 16836 420 17476 8741 9351 7 3 87 3 0
> 3 0 1782060 6943292 5692 128860 152 8252 2116 8536 4025 7673 5 2 88 6 0
> 1 0 1787452 6952868 5676 124644 56 5940 832 7672 3643 8052 6 2 88 5 0
> 0 0 1796428 6964452 5552 122140 0 8956 252 9124 5157 10080 5 3 91 1 0
> 0 1 1798476 6965416 5540 122844 20 1996 436 4436 3824 8384 5 3 91 1 0
> 2 0 1798476 6960840 5548 122828 16 0 212 4508 3559 9714 6 3 88 3 0
> 1 0 1798476 6960744 5556 122844 0 0 0 2408 4049 9925 7 3 89 1 0
> 0 0 1798476 6964672 5556 122932 0 0 164 620 4447 10341 4 4 91 1 0
> 0 1 1818228 6981740 5244 125548 8 19624 200 24472 10605 10163 6 4 86 5 0
> 1 1 1826304 6990164 4716 123008 4 8132 2156 10976 8670 14551 3 3 86 8 0
> 0 0 1829252 6992928 4544 121448 0 2992 1536 5508 4952 14444 3 4 86 7 0
> procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
> r b swpd free buff cache si so bi bo in cs us sy id wa st
> 3 0 1829252 6999360 4544 121448 0 0 0 0 3997 9792 3 4 92 0 0
> 0 0 1829252 6999092 4552 121568 0 0 192 172 3806 9246 4 4 91 0 0
> 1 0 1829252 6996960 4552 121576 0 0 0 0 4103 9302 4 4 92 0 0
> 2 0 1829252 6998984 4556 116736 0 40 4 212 4696 10466 3 5 91 1 0
> 1 0 1829252 6998404 4564 116904 184 0 312 36 4130 10487 3 3 93 1 0
> 0 0 1828228 7003720 4564 116992 1344 0 2736 6312 5595 17908 3 3 86 8 0
> 1 0 1828228 6998980 4572 117596 0 0 184 28 4347 10927 4 4 92 1 0
> 0 1 1827716 7000768 4572 117612 892 0 1484 18276 3538 9155 4 2 86 9 0
> 0 1 1826692 6992004 4572 119228 2404 0 3404 13100 3532 10196 7 1 80 12 0
> 1 0 1825156 6996056 4572 119540 1804 0 1876 15272 4086 13877 6 2 79 13 0
> 1 1 1823364 6994116 4580 119752 1812 0 2004 1840 3567 9753 6 2 79 14 0
> 0 2 1821828 6990780 4588 119816 1248 0 1648 48244 2728 7815 4 1 87 8 0
> 0 1 1818500 6984708 4588 120392 4020 0 4252 0 3833 9440 5 1 84 10 0
> 1 1 1816708 6987868 4588 120972 2604 0 2760 44032 2909 7877 5 1 86 8 0
> 0 1 1813892 6980520 4588 121300 2884 0 2948 38200 2803 7501 4 1 86 9 0
> 0 1 1811588 6982916 4596 121320 2132 0 2332 22492 2572 7048 8 1 84 6 0
> 0 3 1809284 6976112 4596 121480 2556 0 3004 16036 2820 7733 6 1 82 12 0
> 0 1 1806980 6968892 4596 122200 2904 0 3068 18384 4003 12307 4 2 81 13 0
> 0 1 1805188 6973156 4604 122428 2068 0 20068 12 2587 5901 1 1 87 11 0
> 1 2 1802116 6964912 4604 123264 4144 0 15640 72 3851 8710 2 1 82 14 0
> 1 0 1801092 6961472 4612 123680 1056 0 1720 328 3279 6709 6 1 86 6 0
> procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
> r b swpd free buff cache si so bi bo in cs us sy id wa st
> 1 0 1801092 6964964 4612 124068 0 0 712 1100 3987 9326 6 3 89 2 0
> 0 1 1801092 6962704 4612 124152 100 0 2116 1684 3718 10004 3 2 91 4 0
> 0 1 1801092 6961856 4612 124776 684 0 2576 5868 2439 6601 5 1 87 7 0
> 2 0 1801092 6963972 4620 125092 16 0 728 1816 2385 7271 2 2 88 8 0
> 0 0 1801092 6961656 4620 125048 0 0 0 0 3932 10045 4 3 93 0 0
> 2 0 1801092 6963992 4628 125048 0 0 384 6220 4780 12670 4 4 90 2 0
> 0 0 1801092 6963608 4628 125052 0 0 48 0 4342 10433 4 4 92 0 0
> 0 0 1801092 6963588 4628 125164 0 0 64 0 4347 11279 4 4 92 0 0
> 0 1 1800836 6961732 4636 125168 160 0 724 9632 3832 9534 4 4 89 3 0
> 1 0 1800324 6959144 4636 125356 480 0 2928 492 2661 6953 3 1 89 7 0
> 0 2 1796996 6957780 4644 125436 3496 0 4140 19848 2998 9266 6 2 84 8 0
> 0 1 1796228 6962888 4644 125484 668 0 1180 5020 3280 10379 6 2 81 12 0
> 0 1 1795972 6958552 4644 125668 444 0 1208 13716 3183 12078 2 2 88 8 0
> 1 1 1794948 6955404 4656 125808 1180 0 1312 5004 2966 11573 3 2 83 12 0
> 1 0 1794692 6958740 4656 126056 508 0 1324 4576 2738 6384 5 1 88 7 0
> 1 0 1794436 6953764 4664 126676 780 0 1952 1136 2625 5898 8 1 84 7 0
> 1 0 1794436 6955744 4664 126968 32 0 892 1320 3097 7727 7 2 88 4 0
> 0 0 1794436 6957320 4664 126968 4 0 4 16 3637 8272 5 3 92 0 0
> 2 0 1794436 6957408 4664 126972 0 0 84 0 3850 9745 4 4 92 0 0
> 1 0 1794436 6951244 4672 127064 8 0 80 4640 3333 9319 9 3 87 2 0
> 0 1 1794436 6956876 4672 127068 28 0 1452 1468 3919 9485 6 2 87 5 0
>
prev parent reply other threads:[~2018-06-01 8:09 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-30 10:04 Mike Galbraith
2018-06-01 8:09 ` Vlastimil Babka [this message]
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=4505a137-b4e1-4c2a-67d0-3d526fac0503@suse.cz \
--to=vbabka@suse.cz \
--cc=efault@gmx.de \
--cc=hughd@google.com \
--cc=ikalvachev@gmail.com \
--cc=linux-mm@kvack.org \
/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