From: Markham, Richard (RMarkham@hafeleamericas.com)
Date: Thu Apr 18 2002 - 12:43:51 EDT
thanks considerably for all the responses, I wont list names, but you
know who you are. Again the list has proved priceless to shed light
on anything I've ever needed.
key document: the Solaris Memory System
http://www.sun.com/sun-on-net/performance/vmsizing.pdf
:
given below the top information says I had 2254M swap in use.
that's interpreted as the system is using 2gb of the 8gb of VM
on the machine (8VM = 4gb RAM + 4gb diskswap).
The `swap -l` output shows I have no diskswap in use.
swapfile dev swaplo blocks free
/dev/vx/dsk/swapvol 159,8 16 8389632 8389632
another thing to consider is that sleeping processes will be
paged out until they are used again. the document was really
good in that explained the different types of swapping
(ie, hard/soft).
we've had a few complaints about the instance being slow so hence the
magnifying glass on memory usage. the above doc talked about Direct I/O
(the Veritas equivalent, I believe is 'Quick I/O')
which explained some things that are definitely worth considering.
Thanks again.
-----Original Message-----
From: Markham, Richard [mailto:RMarkham@hafeleamericas.com]
Sent: Wednesday, April 17, 2002 3:29 PM
To: Sunman (E-mail)
Subject: ex. top in oracle environment
Below I have put the top output of a system running an oracle database
with two instances. At this time there was only a handful of users in
either of the instances total and most are relatively idle. Noone at
the time was doing any heavy transations or loads.
My question is that with the above given is it normal to have over
2gb of swap in use? Comments and suggestions for things to check would
be greatly appreciated. Thank you.
********-----------------------------------------------------********
last pid: 6940; load averages: 2.07, 2.74, 2.73
15:18:35
263 processes: 261 sleeping, 2 on cpu
CPU states: 39.1% idle, 51.3% user, 8.7% kernel, 0.9% iowait, 0.0% swap
Memory: 4096M real, 962M free, 2254M swap in use, 4939M swap free
PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU COMMAND
3308 oradev 1 40 0 575M 556M sleep 243:09 13.65% oracle
3433 oratest 1 22 0 647M 631M cpu/1 181:12 11.29% oracle
3284 oradev 1 0 0 575M 555M sleep 139:53 5.12% oracle
3391 oratest 11 0 0 648M 631M sleep 45:57 2.67% oracle
6725 oradev 1 0 0 574M 553M sleep 1:04 1.23% oracle
6921 oradev 1 0 0 574M 552M sleep 0:07 1.23% oracle
6804 oradev 1 0 0 574M 553M sleep 0:36 1.19% oracle
6867 oradev 1 10 0 574M 552M sleep 0:16 1.18% oracle
6833 oradev 1 0 0 574M 553M sleep 0:27 0.99% oracle
6495 oratest 1 59 0 647M 628M sleep 1:31 0.96% oracle
6764 oratest 1 0 0 646M 627M sleep 0:28 0.93% oracle
6605 oratest 1 59 0 646M 627M sleep 1:06 0.79% oracle
3445 oratest 1 0 0 647M 629M sleep 4:37 0.77% oracle
3483 oratest 1 0 0 647M 629M sleep 4:36 0.74% oracle
6642 oratest 1 59 0 646M 627M sleep 0:56 0.63% oracle
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:24:13 EDT