Basically, the memory resource control for zone works. But it does not
constrain the physical memory well.
# prstat -Z
ZONEID NPROC SWAP RSS MEMORY TIME CPU ZONE
0 57 240M 310M 1.9% 0:04:03 0.1% global
11 32 162M 235M 1.4% 0:01:08 0.0% bighead
The max-rss of zone bighead is 1G:
zonecfg:bighead> info capped-memory
capped-memory:
physical: 1G
[swap: 7G]
In zone bighead, total memory of an Oracle database is 2 Gb (1.5G SGA + 0.5G
GA). Start up the database in the zone.
# vmstat -p 5
memory page executable anonymous filesystem
swap free re mf fr de sr epi epo epf api apo apf fpi fpo fpf
38894736 15251056 0 0 0 0 0 0 0 0 0 0 0 0 0 0
37248552 13521632 210 797 0 0 0 0 0 0 0 0 0 0 0 0
37236952 13596320 193 126 0 0 0 28574 0 0 1462 0 0 1125 0 0
37212448 13753288 56 217 2 0 0 28807 0 0 1359 0 0 342 2 2
37194568 13700880 78 374 0 0 0 41898 0 0 1303 0 0 420 0 0
# prstat -Z
ZONEID NPROC SWAP RSS MEMORY TIME CPU ZONE
0 62 253M 323M 2.0% 0:05:25 1.5% global
12 55 2100M 1890M 12% 0:01:29 0.4% bighead
The total physical memory drop shown by vmstat is 1.5G(15251056-13700880)
and is 1.6G(1890M-235M) by prstat. But zonecfg limited it to 1G.
Any reason?
Thanks
Victor
This message posted from opensolaris.org
_______________________________________________
opensolaris-discuss mailing list
[email protected]
