From: abdellah.djebli@SICKKIDS.CA
Date: Fri May 23 2003 - 11:46:03 EDT
After setting the vmtune to (-p10 -P30)the paging is gone and the system is
doing a lot better.
The IO like you said, could be network related, which I hgave to
investigate,
because the system has two NICs (10/100), each on a different network.
I have also noticed that aio on the sytem is not used. It is in a define
state
only. Would it help to set it up with 10min/20max aioservers.
Thanks.
Abdu
NotesMail02:~ # lsps -a
Page Space Physical Volume Volume Group Size %Used Active Auto
Type
paging00 vpath0 mail02vg 2944MB 3 yes no
lv
hd6 hdisk0 rootvg 2944MB 6 yes yes
lv
kthr memory page faults cpu
----- ----------- ------------------------ ------------ -----------
r b avm fre re pi po fr sr cy in sy cs us sy id wa
1 1 711869 378 0 0 0 388 1358 0 4475 17748 8384 18 4 51 27
0 2 711869 689 0 0 0 902 3493 0 4640 36422 8052 28 5 26 41
1 1 711869 653 0 0 0 771 3719 0 4621 47462 8295 20 4 38 38
0 2 711875 254 0 0 0 0 0 0 4453 97337 8811 10 11 68 12
1 1 711876 709 0 0 0 514 2384 0 4434 53794 8491 14 6 64 15
2 1 711876 357 0 0 0 513 2869 0 4587 50692 8833 10 7 58 26
1 0 711876 393 0 0 0 513 2921 0 4575 38851 8761 14 4 55 28
0 0 711876 255 0 0 0 0 0 0 4316 37153 8425 8 3 83 6
0 1 711877 426 0 0 0 385 1166 0 4361 51438 8281 12 5 66 17
1 1 711874 647 0 0 0 1157 5775 0 4672 70201 9117 13 6 35 46
1 1 711877 408 0 0 0 386 1252 0 4526 37772 9859 14 6 58 22
3 1 711877 386 0 1 0 513 2557 0 5165 20457 9148 4 6 29 60
0 1 711877 611 0 0 0 642 3095 0 4468 72921 8851 8 6 24 61
1 1 711877 669 0 0 0 642 3105 0 4505 62879 8649 17 8 26 50
0 1 711877 291 0 0 0 0 0 0 4314 16527 8727 4 2 64 30
"Green, Simon"
<Simon.Green@EU.A To: aix-l@Princeton.EDU
LTRIA.COM> cc:
Sent by: IBM AIX Subject: Re: performance question
Discussion List
<aix-l@Princeton.
EDU>
05/23/2003 10:06
AM
Please respond to
IBM AIX
Discussion List
With your earlier situation, I would have assumed that the I/O wait was the
result of activity on the disks holding your paging spaces. However, if
you've now eliminated paging that wouldn't be the case. (Of course, that
depends: what do you mean by "hardly doing any paging"?)
IO waits are not necessarily disk: another possibility is network. As this
is a mail server, that seems worth investigating.
Simon Green
Altria ITSC Europe s.a.r.l.
AIX-L Archive at http://marc.theaimsgroup.com/?l=aix-l&r=1&w=2
AIX FAQ at http://www.faqs.org/faqs/aix-faq/
N.B. Unsolicited email from vendors will not be appreciated.
> -----Original Message-----
> From: abdellah.djebli@SICKKIDS.CA [mailto:abdellah.djebli@SICKKIDS.CA]
> Sent: 23 May 2003 13:48
> To: aix-l@Princeton.EDU
> Subject: Re: performance question
>
>
> Yes, I am slowly decreasing them, I am at (-p10 -P40) and I
> can already see
> a major improvement as far as
> paging. The system is hardly doing any paging now. I still
> have to figure
> out why we have the IO wait while
> nmon and/or iostat does not show very busy disks.
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 22:16:51 EDT