Well, I found out more about. As long as the default scheduling class is set,
the system can handle 32 single-thread processes.
Victor
This message posted from opensolaris.org
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
I just tested out that with FSS this system is able to handle 32 processes with
single thread.
Victor
This message posted from opensolaris.org
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
I did a couple of tests, in most of tests, importance could not play its role.
With following configuration,
[EMAIL PROTECTED] zonecfg -z global info dedicated-cpu
dedicated-cpu:
ncpus: 2-20
importance: 4
[EMAIL PROTECTED] zonecfg -z zoneA info dedicated-cpu
dedicated-cpu:
One way to find out how many CPUs are allocated to each zone is prstat -Z based
on CPU%. Is there any better way?
Thanks
Victor
This message posted from opensolaris.org
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
I tried with locked:1G but Oracle opened.
capped-memory:
physical: 1G
[swap: 7G]
[locked: 1G]
Oracle database took for a while to start up.
The RSS size of the zone jumped around
# prstat -Z
ZONEIDNPROC SWAP RSS MEMORY TIME CPU ZONE
36 54 1824M
It is a T2000 server with 8 cores. T1 processor does good for the system
running process with multithreads. It seems to me that it can handle 1.5*number
of core=1.5*8=12 processes with single thread. Is this formula correct for T1
processor?
perl -e 'while (--$ARGV[0] and fork) {}; while ()
Basically, the memory resource control for zone works. But it does not
constrain the physical memory well.
# prstat -Z
ZONEIDNPROC 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
Given the importance of zones and some other info, is there any formula to
calculate the number of CPUs that system will allocate to each zone?
e.g.
Following system has only two zones with dynamic pool service enabled.
Total number of CPU in the system is 32.
# zonecfg -z zoneA info
profile:::tick-1sec
/this-cnt == 10/
{
normalize(@runq, 1);
printa([EMAIL PROTECTED], @runq);
exit(0);
}
The output is an integer, but I want a decimal, e.g. 7.4 instead of 7. How
could I do that?
Thanks
Victor
This message posted from opensolaris.org
There is a file UserPreferencesSettings.jsp in /mail/products/uwc/uwc/common/,
but 'find' can not find it except when the search path is set as
/mail/products/uwc/uwc/common/.
user ls /mail/products/uwc/uwc/common/User*
/mail/products/uwc/uwc/common/UserPreferencesPassword.jsp
Never mind, it is linked in the middle.
# pwd
/mail/products/uwc
# ls -l uwc
lrwxrwxrwx 1 user staff 34 Sep 20 12:01 uwc -
/var/mail/products/uwc/staging/uwc
Victor
This message posted from opensolaris.org
___
opensolaris-discuss
In a ufs volume with simple two-way mirror, after a submirror is detached, the
data in its slice are accessible, i.e. we can mount that slice later, and
access the data. But not for the submirror slice in a zfs pool.
Victor
This message posted from opensolaris.org
Hi, all,
This is brand new box with standard installation. O.S.5.9 9/04. Auto_home is on
in auto_master file. Out of curiosity, why is the user not permitted to go to
/export/home physically?
[EMAIL PROTECTED] cd /export
[EMAIL PROTECTED] ls
download home
[EMAIL PROTECTED] cd home
bash: cd:
was prior to the disk slice being mirrored.
Ours is Solaris 9 9/04. Looking at the commented entry in /etc/vfstab,
#/dev/dsk/c1t0d0s1 - - swap - no -, I think that the machine is installed with
Web Start or something. What is your experience or opinion on his opinion?
Thanks
Victor Feng
Our Solaris boxes were installed by vendor. I observed that:
1. Dump device: /dev/dsk/c1t0d0s1 (dedicated)
2. /dev/dsk/c1t0d0s1 is part of mirror /dev/md/dsk/d1
3. /dev/md/dsk/d1 is the swap
See the following script for details.
Questions:
1. I wonder why the vendor uses /dev/dsk/c1t0d0s1 as dump
Thank you for your warning of savecore -L. I am going to do dumpadm -d swap.
Victor
This message posted from opensolaris.org
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
I have same problem as you do. The smpatch worked for our Solaris 5.9 a week
ago. But now it gives that error message. Have you figured it out?
Victor
This message posted from opensolaris.org
___
opensolaris-discuss mailing list
The smpatch analyze, smpatch download worked last week. Today, they always give
error message response code was 500, and the log message says can not create
cache downloader.
We have four brand new Sun Solaris 9 boxes. Last week two of them were well
patched using smpatch, and today these two
18 matches
Mail list logo