What is the best way to analyze memory usage inside a zone with capped
memory? I have a server with numerous zones all with capped memory. If
I run prtconf inside the zone or top the zone reports total physical
memory in the zone instead of it's capped limit. Also I believe rcapd
in the global zone is working pretty hard; it's used about 305 hours
of cpu time in 68 days.

Thanks for any tips or insight.

Vic

On Jul 15, 2:01=A0pm, "victor.en...@gmail.com" <victor.en...@gmail.com>
wrote:
> What is the best way to analyze memory usage inside a zone with capped
> memory? I have a server with numerous zones all with capped memory. If
> I run prtconf inside the zone or top the zone reports total physical
> memory in the zone instead of it's capped limit. Also I believe rcapd
> in the global zone is working pretty hard; it's used about 305 hours
> of cpu time in 68 days.
>
> Thanks for any tips or insight.
>
> Vic

Looking into rcapstat now...

On Jul 15, 7:17=A0pm, Vic <victor.en...@gmail.com> wrote:
> On Jul 15, 2:01=A0pm, "victor.en...@gmail.com" <victor.en...@gmail.com>
> wrote:
>
> > What is the best way to analyze memory usage inside a zone with capped
> > memory? I have a server with numerous zones all with capped memory. If
> > I run prtconf inside the zone or top the zone reports total physical
> > memory in the zone instead of it's capped limit. Also I believe rcapd
> > in the global zone is working pretty hard; it's used about 305 hours
> > of cpu time in 68 days.
>
> > Thanks for any tips or insight.
>
> > Vic
>
> Looking into rcapstat now...

prstat in the non global zone will show memory usage. zonecfg will
show the capped memory size. prstat with the -Z option will show about
zones.

# prstat -Z
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/
NLWP
   741 root     8704K 5744K sleep   59    0  10:55:27 6.0% svc.configd/
14
   763 root     8448K 6560K wait     9    0  10:59:37 6.0% svc.configd/
14
 12968 root     4128K 3504K cpu1    59    0   0:00:00 0.7% prstat/1
   739 root     9280K 7344K wait    48    0   2:13:00 0.4% svc.startd/
16
   761 root     9248K 7072K sleep   59    0   2:11:15 0.4% svc.startd/
18
 12780 root     4088K 3472K sleep   59    0   0:00:00 0.1% prstat/1
     1 root     2912K 1184K sleep   59    0   0:09:49 0.1% init/1
  3021 noaccess  214M  134M sleep   59    0   0:10:44 0.1% java/17
 13137 root     1664K 1264K wait    48    0   0:00:00 0.1% svc-snmpdx/
1
  2998 noaccess  218M  138M sleep   59    0   0:10:34 0.1% java/17
  2569 root      214M  136M sleep   59    0   0:09:37 0.1% java/17
   683 root       25M   15M sleep   59    0   0:03:01 0.0% Xsun/1
   942 root     6296K 4160K sleep   59    0   0:00:20 0.0% nscd/28
  1020 root     6360K 4016K sleep   59    0   0:02:13 0.0% nscd/29
   729 root       13M 2672K sleep   59    0   0:02:12 0.0% dtgreet/1
   154 root     7000K 4496K sleep   59    0   0:00:25 0.0% nscd/31
  1346 postgres   49M   38M sleep   59    0   0:01:42 0.0% postgres/1
  1828 root     4152K 1816K sleep   59    0   0:01:54 0.0% syslogd/13
  1745 root     4216K 2272K sleep   59    0   0:01:51 0.0% syslogd/13
   907 root     6192K 3568K sleep   59    0   0:01:16 0.0% nscd/29
  9899 unix     2704K 2024K sleep   59    0   0:00:00 0.0% telnet/1
  1046 root     3104K 1712K sleep   59    0   0:00:51 0.0% xntpd/1
  1022 root     3104K 1760K sleep   59    0   0:00:48 0.0% xntpd/1
  1179 root     3064K 1576K sleep   59    0   0:00:51 0.0% xntpd/1
   233 root     3104K 1448K sleep  100    -   0:00:42 0.0% xntpd/1
   752 root     8496K 3568K sleep   59    0   0:00:36 0.0% svc.configd/
13
   357 root     1664K  768K sleep   59    0   0:00:00 0.0% utmpd/1
   512 root     4264K 1960K sleep   59    0   0:00:00 0.0% syslogd/13
   333 root     7040K 4800K sleep   59    0   0:00:10 0.0% inetd/4
   329 root     2512K 1080K sleep   59    0   0:00:00 0.0% sac/1
ZONEID    NPROC  SWAP   RSS MEMORY      TIME  CPU ZONE
     1       33  211M  259M    17%  13:25:21 6.6% ATPUGSAD
     2       29  182M  238M    16%  13:26:56 6.5% NAUEEREEFS
     0       52   94M  110M   7.1%   0:18:33 0.9% global
     3       34  183M  244M    16%   0:12:06 0.2% AALOROTOM


Total: 148 processes, 575 lwps, load averages: 3.28, 3.27, 3.25


Kartik Vashishta

On Wed, 15 Jul 2009 11:01:57 -0700, victor.engle@gmail.com wrote:

> What is the best way to analyze memory usage inside a zone with capped
> memory? I have a server with numerous zones all with capped memory. If I
> run prtconf inside the zone or top the zone reports total physical
> memory in the zone instead of it's capped limit. Also I believe rcapd in
> the global zone is working pretty hard; it's used about 305 hours of cpu
> time in 68 days.
> 
> Thanks for any tips or insight.
> 
> Vic

If you're on Solaris 10u7 or later the zone will report the correct value 
for its total memory in top if you're using capped-memory physical value.

Using rcapstat -z 1 or something like that from the globalzone will give 
the rcap activity.  Also look at prstat -Z