Re:_FYI;_14.3:_A_discord_report_o f_Wired_Memory_growing_to_17_GiBy tes_over_something_like_60_days;_ ARC_shrinks_to,_say,_1942_MiBytes

i ran this oneliner sequence. it was a (poor) attempt to force push ram empty. it mostly got empty. but. i think i have a problem here. do you also see problem here? i have 532.07 mb of kernel memory going somewhere. i won't be looking if it doesn't get slower or "overflow" in rare cases. is that a zfs write cache? note that this is on 13.5-RELEASE-p3, currently no other releases, stables or currents where i can perform same tests as a comparison, nor any more powerful systems. i'm reading here that zfs can't give ram back if it's middle of "something". unsure what to think


04:32,ketas@green:~> stress-ng --vm 1 --vm-bytes 100% -t 10s ; top -b | head -7 ; zf
s-stats -a | fgrep 'Kernel Memory:'
stress-ng: info:  [61750] setting to a 10 secs run per stressor
stress-ng: info:  [61750] dispatching hogs: 1 vm
stress-ng: info:  [61814] vm: using 3.88G per stressor instance (total 3.88G of 3.88
G available memory)
stress-ng: warn:  [61750] metrics-check: all bogo-op counters are zero, data may be
incorrect
stress-ng: info:  [61750] skipped: 0
stress-ng: info:  [61750] passed: 1: vm (1)
stress-ng: info:  [61750] failed: 0
stress-ng: info:  [61750] metrics untrustworthy: 0
stress-ng: info:  [61750] successful run completed in 45.04 secs
last pid: 62582;  load averages:  0.57,  0.92,  1.06  up 2+01:45:26    04:34:19
261 processes: 1 running, 260 sleeping
CPU:  6.4% user,  2.6% nice,  8.1% system,  0.6% interrupt, 82.2% idle
Mem: 93M Active, 20M Inact, 484K Laundry, 1246M Wired, 2492M Free
ARC: 392M Total, 156M MFU, 102M MRU, 2794K Anon, 3493K Header, 127M Other
     38M Compressed, 224M Uncompressed, 5.88:1 Ratio
Swap: 16G Total, 2627M Used, 13G Free, 16% Inuse
Kernel Memory:                                  321.93  MiB
04:34,ketas@green:~>