[Bug 290207] [ZFS] lowering "vfs.zfs.arc.max" to a low value causes kernel threads of "arc_evict" to use 91% CPU and disks to wait. System gets unresponsive...
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Fri, 31 Oct 2025 13:56:36 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=290207
--- Comment #19 from Nils Beyer <nbe@vkf-renzel.de> ---
(In reply to Mark Millard from comment #18)
======
before
======
right after reboot - no Xorg started, just plain tty:
# top | head -7
last pid: 3829; load averages: 0.12, 0.03, 0.01; battery: 99%
up
0+00:00:25 14:49:10
538 threads: 13 running, 471 sleeping, 54 waiting
CPU: 0.4% user, 0.0% nice, 1.8% system, 0.1% interrupt, 97.7% idle
Mem: 35M Active, 41M Inact, 539M Wired, 13G Free
ARC: 93M Total, 18M MFU, 71M MRU, 256K Anon, 576K Header, 1957K Other
58M Compressed, 126M Uncompressed, 2.16:1 Ratio
Swap: 16G Total, 16G Free
# vmstat -o | sort -nr -k5,5 -k1,1 -k10,10 -k9,9 | grep -Ev '[0-9]+ +[0-9]+
+[0-9]+ +[0-9]+ 0 ' | head
2408 0 0 0 2408 1 0 WB ph
1978 0 0 0 1978 3 0 WB ph
200 0 0 0 200 3 0 WB ph
148 0 0 0 148 3 0 WB ph
101 0 0 0 101 3 0 WB ph
68 0 0 0 68 3 0 WB ph
66 0 0 0 66 1 0 WB ph
51 0 0 0 51 3 0 WB ph
47 0 0 0 47 3 0 WB ph
45 0 0 0 45 3 0 WB ph
# vmstat -o | sort -nr -k5,5 -k1,1 -k10,10 -k9,9 | grep -Ev '[0-9]+ +[0-9]+
+[0-9]+ +[0-9]+ 0 ' | tail
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
RES ACT INACT LAUND WIRED REF SHD CM TP PATH
# vmstat -o | sort -nr -k5,5 -k1,1 -k10,10 -k9,9 | grep -Ev '[0-9]+ +[0-9]+
+[0-9]+ +[0-9]+ 0 ' | wc -l
52
======
action
======
recursively deleting a filled "/usr/src" (resp. "/usr/src.test" which is a
pre-made copy done before the reboot)
# rm -rf /usr/src.test
=====
after
=====
# top | head -7
last pid: 4166; load averages: 0.00, 0.01, 0.00; battery: 99% up
0+00:04:26 14:53:10
535 threads: 13 running, 468 sleeping, 54 waiting
CPU: 0.0% user, 0.0% nice, 0.3% system, 0.0% interrupt, 99.6% idle
Mem: 34M Active, 42M Inact, 1634M Wired, 12G Free
ARC: 114M Total, 13M MFU, 94M MRU, 256K Anon, 1179K Header, 5782K Other
94M Compressed, 161M Uncompressed, 1.72:1 Ratio
Swap: 16G Total, 16G Free
# vmstat -o | sort -nr -k5,5 -k1,1 -k10,10 -k9,9 | grep -Ev '[0-9]+ +[0-9]+
+[0-9]+ +[0-9]+ 0 ' | head
2440 0 0 0 2440 1 0 WB ph
1978 0 0 0 1978 3 0 WB ph
200 0 0 0 200 3 0 WB ph
148 0 0 0 148 3 0 WB ph
101 0 0 0 101 3 0 WB ph
68 0 0 0 68 3 0 WB ph
66 0 0 0 66 1 0 WB ph
51 0 0 0 51 3 0 WB ph
47 0 0 0 47 3 0 WB ph
45 0 0 0 45 3 0 WB ph
# vmstat -o | sort -nr -k5,5 -k1,1 -k10,10 -k9,9 | grep -Ev '[0-9]+ +[0-9]+
+[0-9]+ +[0-9]+ 0 ' | tail
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
3 0 0 0 3 3 0 WB ph
RES ACT INACT LAUND WIRED REF SHD CM TP PATH
# vmstat -o | sort -nr -k5,5 -k1,1 -k10,10 -k9,9 | grep -Ev '[0-9]+ +[0-9]+
+[0-9]+ +[0-9]+ 0 ' | wc -l
52
--
You are receiving this mail because:
You are the assignee for the bug.