From: Mike Simons <msimons@moria.simons-clan.com>
To: Linux Memory Management List <linux-mm@kvack.org>
Subject: 2.3.99-pre8 + riel patch 2, results.
Date: Mon, 15 May 2000 19:14:30 -0400 [thread overview]
Message-ID: <20000515191430.A5677@moria.simons-clan.com> (raw)
Hi all,
This info plus the exact patches used are available at:
http://moria.simons-clan.com/~msimons/
2.3.99-pre8 plus "riel-patch2.diff"
===================================
vmstat 1 | tee log ... running in one window.
mmap002 ... running in another window.
All tests done in multi-user console mode with nothing else happening.
First run the cache fills up... swpd starts up. Two seconds after swpd
mmap002 gets killed. The wierd-ish thing is vmstat shows only "bi" up
until the time swap started...
- why wasn't there any bo?
The system becomes very unresponsive just _after_ killing mmap002.
then the machine lags badly for about 6 seconds (which show up as three
vmstat lines) until that point things were fine and after that it's fine.
I can run mmap002 a few more times... but now that the cache is full
the system will kill mmap002 very quickly...
Second run it takes 7 seconds to kill mmap002.
Third run takes 4 seconds to kill mmap002, and 2 seconds later init dies.
The wierd part is that when init dies the whole system is locked... not
even caps-lock light works. In the past (2.2.*) when init was killed
the system would continue to function for at least several minutes when
very little else was happening.
A few times the system kills a whole list of processes before locking:
killing process mmap002, klogd, init, init, sysklogd, sysklogd,
vmstat, bash, tee, bash, init, init, init, then _solid_ hang...
This within seconds of starting of the third mmap002 run.
Hope that helps,
Mike Simons
first run of mmap002
procs memory swap io system cpu
r b w swpd free buff cache si so bi bo in cs us sy id
1 0 0 0 16340 85212 8516 0 0 1677 3 294 384 4 4 92
0 0 0 0 16316 85212 8520 0 0 0 0 108 9 1 0 99
0 0 0 0 16316 85212 8520 0 0 0 0 106 15 0 0 100
0 0 0 0 16312 85216 8520 0 0 1 0 109 15 0 0 100
0 0 0 0 16304 85216 8520 0 0 0 0 107 13 0 0 100
0 0 0 0 16304 85216 8520 0 0 0 0 105 11 0 0 100
0 0 0 0 16304 85216 8520 0 0 0 27 124 11 0 0 100
0 1 0 0 16220 85216 8524 0 0 1 0 102 12 0 0 100
0 1 0 0 15712 76436 18608 0 0 2646 0 350 433 3 2 95
1 0 0 0 14976 71420 24816 0 0 1554 0 206 212 4 3 93
1 0 0 0 15056 65536 31152 0 0 1585 0 204 216 5 2 93
0 1 0 0 14484 60552 37168 0 0 1506 0 199 209 2 1 97
1 0 0 0 14572 54732 43440 0 0 1569 0 202 212 5 4 91
0 1 0 0 9320 53588 49776 0 0 1586 0 340 482 3 0 97
0 1 0 0 2836 53592 56048 0 0 1569 0 258 318 4 0 96
1 0 0 0 19924 32896 62128 0 0 1522 0 200 204 3 7 90
0 1 0 0 13368 32900 68464 0 0 1585 0 205 211 1 2 97
1 0 0 0 7016 32908 74608 0 0 1538 0 224 206 2 1 97
0 1 0 0 2464 31248 80816 0 0 1553 0 204 209 1 4 95
1 0 0 0 31904 220 86316 0 0 1394 4 212 203 3 7 90
1 0 0 0 25932 248 92076 0 0 1447 0 207 213 1 2 97
1 2 0 0 19888 344 97732 0 0 1438 0 218 222 6 0 94
1 0 0 0 14000 376 103492 0 0 1448 0 218 207 5 0 95
1 0 0 0 7848 380 109444 0 0 1489 0 197 196 6 1 93
0 1 0 0 2900 308 114452 0 0 1426 7 322 375 4 2 94
1 0 1 112 3180 72 117816 0 112 1042 61 336 375 3 13 84
2 0 0 112 3056 68 118032 0 0 1157 146 487 645 4 6 90
0 1 0 720 2120 96 119648 0 608 891 322 464 375 3 18 79
VM: killing process mmap002
0 1 1 936 2668 72 119360 112 264 499 1102 449 119 0 39 61
1 1 1 1368 2216 120 119536 8 436 60 12818 7676 13356 0 7 93
0 2 0 1672 2576 116 120104 320 312 248 15210 7575 11462 0 9 91
0 0 0 1672 2424 124 120236 28 0 35 0 268 338 0 0 100
0 0 0 1672 2424 124 120236 0 0 0 0 520 845 0 0 100
0 0 0 1672 2424 124 120236 0 0 0 0 300 405 0 0 100
0 0 0 1672 2424 124 120236 0 0 0 0 132 10 0 0 100
0 0 0 1672 2424 124 120236 0 0 0 0 132 8 0 0 100
r b w swpd free buff cache si so bi bo in cs us sy id
procs memory swap io system cpu
second run of mmap002
procs memory swap io system cpu
r b w swpd free buff cache si so bi bo in cs us sy id
1 0 0 1660 3820 472 118200 1 2 341 40 163 120 1 1 98
0 0 0 1660 3808 480 118204 0 0 2 0 112 13 0 0 100
0 0 0 1660 3808 480 118204 0 0 0 0 103 7 0 0 100
0 1 0 1528 2772 432 119100 0 0 917 3 186 165 17 5 78
1 0 0 1464 3084 136 119068 0 0 846 12 178 141 38 6 56
1 0 0 1600 1824 76 120664 12 136 1006 101 212 166 7 6 87
0 1 3 1592 1324 124 119824 168 36 320 2154 411 330 0 19 81
VM: killing process mmap002
0 2 1 1904 2460 148 118564 156 324 188 17503 9158 16378 0 7 93
0 1 0 1884 2200 176 118760 76 0 61 4565 638 759 0 2 98
0 0 0 1884 2196 176 118760 0 0 0 0 625 834 0 1 99
0 0 0 1884 2188 180 118760 0 0 1 0 538 876 0 1 99
0 0 0 1884 2188 180 118760 0 0 0 0 269 345 0 0 100
0 0 0 1884 2188 180 118760 0 0 0 0 123 10 0 0 100
0 0 0 1884 2188 180 118760 0 0 0 0 132 8 0 0 100
third run of mmap002
procs memory swap io system cpu
r b w swpd free buff cache si so bi bo in cs us sy id
1 0 0 1868 2832 272 117968 2 3 327 74 174 139 1 1 98
0 0 0 1868 2800 276 117972 0 0 1 0 131 51 0 0 100
1 0 0 1904 2540 292 118104 0 36 484 10 156 95 32 9 59
1 0 0 1812 3572 128 118864 0 36 1004 24 207 170 25 10 65
0 1 1 2780 1380 100 121416 484 1016 1101 3432 797 392 1 20 79
0 1 1 3056 2228 92 121160 340 340 109 8295 6189 11292 0 10 90
[...death of system... I'm surprised this made it to disk...]
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux.eu.org/Linux-MM/
reply other threads:[~2000-05-15 23:14 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=20000515191430.A5677@moria.simons-clan.com \
--to=msimons@moria.simons-clan.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