Subject: Re: question about vmstat -m
To: Wojciech Puchar <wojtek@wojtek.from.pl>
From: Jason R Thorpe <thorpej@zembu.com>
List: netbsd-users
Date: 01/24/2001 08:33:37
On Wed, Jan 24, 2001 at 11:53:28AM +0100, Wojciech Puchar wrote:
> a) why "rndsample" needs over half a megabyte memory?!
> b) mclpl what is it and why is it so large
mclpl is "mbuf cluster pool". They're big network bufers.
> c) what is sigapl, procpl
sigapl == "sigaction pool", procpl == "proc pool". sigaction and proc
structures.
> it no problem on 384MB machine but how would it be on 16MB? all pools
> below are 5MB and does not count buffer cache
Well, on a smaller machine (e.g. 16MB RAM), the pagedaemon would come around
and reclaim the "Idle" pages. Those are lazily reclaimed so that the pools
can use them more quickly if they need them.
Also, on a 16MB machine, various other limits are tuned differently such
that the pools don't generally grow as big.
Consider this system (a 32MB AlphaStation 200 that acts as a file server,
mail server, DNS server, KDC, and web server).
Note the "Pgrel" figures, indicating how many times a pool gave a page
back to the free list.
yeah-baby:thorpej 2$ vmstat -m
Kmem statistics are not being gathered by the kernel.
Memory resource pool statistics
Name Size Requests Fail Releases Pgreq Pgrel Npage Hiwat Minpg Maxpg Idle
phpool 80 1413 0 1329 3 2 1 2 0 inf 0
pcgpool 152 372 0 371 1 0 1 1 0 inf 0
pmappl 96 51783 0 51730 1 0 1 1 0 inf 0
l1ptpl 8192 1920 0 1894 494 456 38 46 0 inf 12
pmasnpl 24 51783 0 51730 1 0 1 1 0 inf 0
pmasngenpl 24 51783 0 51730 1 0 1 1 0 inf 0
pvpl 40 17801812 0 17798614 211 187 24 27 1 inf 2
vmsppl 288 51783 0 51730 17 14 3 3 0 inf 0
vmmpepl 96 2785132 0 2784042 148 133 15 18 0 inf 0
uaoeltpl 96 431 0 416 1 0 1 1 0 inf 0
aobjpl 88 1 0 0 1 0 1 1 0 inf 0
amappl 56 631523 0 631042 27 23 4 5 0 inf 0
bufpl 192 1441313 0 1441313 30 29 1 9 0 inf 1
mbpl 256 43870373 0 43870287 272 267 5 8 1 inf 2
mclpl 2048 9269470 0 9269393 766 740 26 45 2 256 6
sockpl 288 475626 0 475231 84 69 15 16 0 inf 0
ttypl 440 34 0 0 2 0 2 2 0 inf 0
rndsample 536 101814 0 101810 1 0 1 1 0 inf 0
procpl 720 52132 0 52068 57 50 7 9 0 inf 0
pgrppl 40 6730 0 6696 1 0 1 1 0 inf 0
pcredpl 32 52132 0 52068 1 0 1 1 0 inf 0
plimitpl 160 5152 0 5143 1 0 1 1 0 inf 0
rusgepl 144 52068 0 52068 161 160 1 1 0 inf 1
filepl 72 3476239 0 3475754 5 0 5 5 0 inf 0
cwdipl 24 52121 0 52068 1 0 1 1 0 inf 0
fdescpl 216 52121 0 52068 11 9 2 3 0 inf 0
pnbufpl 1024 17628133 0 17628133 15 13 2 2 0 inf 2
vnodepl 336 778 0 0 33 0 33 33 0 inf 0
ncachepl 112 778 0 0 11 0 11 11 0 inf 0
sdpcpool 192 243221 0 243220 117 116 1 9 0 inf 0
ffsinopl 256 10950809 0 10950045 25 0 25 25 0 inf 0
lfsinopl 256 0 0 0 0 0 0 0 0 inf 0
cd9660nopl 192 0 0 0 0 0 0 0 0 inf 0
msdosnopl 176 0 0 0 0 0 0 0 0 inf 0
nfsnodepl 376 1 0 0 1 0 1 1 0 inf 0
nfsvapl 160 1 0 0 1 0 1 1 0 inf 0
extent 40 258013 0 257306 14 10 4 6 0 inf 0
scxspl 208 1516617 0 1516617 4 3 1 1 0 inf 1
wdcspl 88 4999375 0 4999375 37 36 1 1 0 inf 1
rtentpl 248 492 0 396 4 0 4 4 0 inf 0
rttmrpl 64 0 0 0 0 0 0 0 0 inf 0
ipqepl 72 28944 0 28944 164 163 1 1 0 inf 1
ipflowpl 112 0 0 0 0 0 0 0 0 inf 0
inpcbpl 144 328498 0 328392 23 21 2 3 0 inf 0
tcpcbpl 248 22446 0 22378 10 7 3 3 0 inf 0
synpl 216 17330 0 17330 108 107 1 1 0 inf 1
sigapl 1064 52121 0 52068 80 70 10 12 0 inf 0
swp vnx 32 0 0 0 0 0 0 0 0 inf 0
swp vnd 200 0 0 0 0 0 0 0 0 inf 0
In use 1560K, total allocated 2088K; utilization 74.8%
yeah-baby:thorpej 3$
--
-- Jason R. Thorpe <thorpej@zembu.com>