Board index Aix

Paging 2500/sec a problem

Paging 2500/sec a problem

Post by David Kropma » Wed, 15 Sep 1999 04:00:00



HI

We have a utility that is reporting 2500 pages/sec in. Is that a problem and
what utility should I be using to determine the actual value.

What is a good benchmark for paging.

Thanks

David


 
 
 
Top

Paging 2500/sec a problem

Post by ArctcCr » Wed, 15 Sep 1999 04:00:00


>HI

>We have a utility that is reporting 2500 pages/sec in. Is that a problem and
>what utility should I be using to determine the actual value.

>What is a good benchmark for paging.

>Thanks

>David



Is that pages in or pages out?  Or both?

 
 
 
Top

Paging 2500/sec a problem

Post by maarten kreuge » Wed, 15 Sep 1999 04:00:00



> HI

> We have a utility that is reporting 2500 pages/sec in. Is that a problem and
> what utility should I be using to determine the actual value.

> What is a good benchmark for paging.

> Thanks

> David



What you are seeing is just ordinary IO. In AIX, all IO is done by the
paging algorithm. If you read a file, the kernel sort of 'assumes' the
page the file is on (which is located on a filesystem) is already in
memory. It probably is not, so a page fault is generated and the page is
swapped in, and your file is available. One page in.

So if you see 2500 pages/sec in this will add up to 10MB of disk reads.

If you want to see the pagingspace IO, check vmstat. If you see high
numbers (>10) continuously in the 'pi' or the 'po' column, your system
is paging heavily.

Maarten Kreuger         AIX Services            IBM The Netherlands

 
 
 
Top

Paging 2500/sec a problem

Post by Norman Levi » Wed, 15 Sep 1999 04:00:00



> HI

> We have a utility that is reporting 2500 pages/sec in. Is that a problem and
> what utility should I be using to determine the actual value.

vmstat has a po/pi column.  I'm curious what vmstat is saying.
--
Norman Levin
vm/dynAmIX inc.
 
 
 
Top

Paging 2500/sec a problem

Post by David Kropma » Sat, 18 Sep 1999 04:00:00


It is pages in.


>>HI

>>We have a utility that is reporting 2500 pages/sec in. Is that a problem
and
>>what utility should I be using to determine the actual value.

>>What is a good benchmark for paging.

>>Thanks

>>David


>Is that pages in or pages out?  Or both?

 
 
 
Top

Paging 2500/sec a problem

Post by David Kropma » Sat, 18 Sep 1999 04:00:00


This is the result of vmstat. Do you think we have a problem?

prd:/home/trdkr>vmstat 3 5
kthr     memory             page              faults        cpu
----- ----------- ------------------------ ------------ -----------
 r  b   avm   fre  re  pi  po  fr   sr  cy  in   sy  cs us sy id wa
 0  0 236418 66078   0   3   3 168  432   0 176  159 336 27  8 41 23
 6  3 280844 20899   0  12   0   0    0   0 703 12551 4224 65 32  0  3
 7  4 312397     4   0  14 440 3767 9374   0 878 10707 4827 66 28  0  6
 4  5 317819   189   0  15 943 1920 4996   0 903 8282 4637 47 17  0 36
 9  3 318859     0   0  46  62 782 24175   0 834 11210 5286 77 22  0  2



>> HI

>> We have a utility that is reporting 2500 pages/sec in. Is that a problem
and
>> what utility should I be using to determine the actual value.

>vmstat has a po/pi column.  I'm curious what vmstat is saying.
>--
>Norman Levin
>vm/dynAmIX inc.

 
 
 
Top

Paging 2500/sec a problem

Post by kcras » Sat, 18 Sep 1999 04:00:00


Yeah there is problem.  Many...I am curious what you are
running on this box.  The big one that sticks out that
would point to a lot of things is the 7-9 processes in the
run queue.  But yes, this box needs some tuning and/or
another box next to it.

* Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
The fastest and easiest way to search and participate in Usenet - Free!

 
 
 
Top

Paging 2500/sec a problem

Post by Geral » Sun, 26 Sep 1999 04:00:00


What was running approximately 6-12 seconds after you started vmstat?
Whatever it is is chewing up all your memory. You are memory starved.
The system had to steal memory pages from "peter" to give them to
"paul".

Under the same conditions, can you also run "lsps -a" and "iostat 3 5"

How much physical memory do you have? Also, it looks like you are CPU
bound too, since your user+system percentage is consistently above
80%.


> This is the result of vmstat. Do you think we have a problem?

> prd:/home/trdkr>vmstat 3 5
> kthr     memory             page              faults        cpu
> ----- ----------- ------------------------ ------------ -----------
>  r  b   avm   fre  re  pi  po  fr   sr  cy  in   sy  cs us sy id wa
>  0  0 236418 66078   0   3   3 168  432   0 176  159 336 27  8 41 23
>  6  3 280844 20899   0  12   0   0    0   0 703 12551 4224 65 32  0  3
>  7  4 312397     4   0  14 440 3767 9374   0 878 10707 4827 66 28  0  6
>  4  5 317819   189   0  15 943 1920 4996   0 903 8282 4637 47 17  0 36
>  9  3 318859     0   0  46  62 782 24175   0 834 11210 5286 77 22  0  2



> >> HI

> >> We have a utility that is reporting 2500 pages/sec in. Is that a problem
> and
> >> what utility should I be using to determine the actual value.

> >vmstat has a po/pi column.  I'm curious what vmstat is saying.
> >--
> >Norman Levin
> >vm/dynAmIX inc.

 
 
 
Top

Paging 2500/sec a problem

Post by Dmitry Gordo » Tue, 28 Sep 1999 04:00:00


IMHO we definetly have 2 problems:
1. CPU - High r and b columns ratio ( more details with sar -q )
2. Memory - may be as a result of 1. pagin/pageout rate is getting high. (more
details with svmon)
It seems that you ran many processes at once.

Good luck
Dmitry


> What was running approximately 6-12 seconds after you started vmstat?
> Whatever it is is chewing up all your memory. You are memory starved.
> The system had to steal memory pages from "peter" to give them to
> "paul".

> Under the same conditions, can you also run "lsps -a" and "iostat 3 5"

> How much physical memory do you have? Also, it looks like you are CPU
> bound too, since your user+system percentage is consistently above
> 80%.


> > This is the result of vmstat. Do you think we have a problem?

> > prd:/home/trdkr>vmstat 3 5
> > kthr     memory             page              faults        cpu
> > ----- ----------- ------------------------ ------------ -----------
> >  r  b   avm   fre  re  pi  po  fr   sr  cy  in   sy  cs us sy id wa
> >  0  0 236418 66078   0   3   3 168  432   0 176  159 336 27  8 41 23
> >  6  3 280844 20899   0  12   0   0    0   0 703 12551 4224 65 32  0  3
> >  7  4 312397     4   0  14 440 3767 9374   0 878 10707 4827 66 28  0  6
> >  4  5 317819   189   0  15 943 1920 4996   0 903 8282 4637 47 17  0 36
> >  9  3 318859     0   0  46  62 782 24175   0 834 11210 5286 77 22  0  2



> > >> HI

> > >> We have a utility that is reporting 2500 pages/sec in. Is that a problem
> > and
> > >> what utility should I be using to determine the actual value.

> > >vmstat has a po/pi column.  I'm curious what vmstat is saying.
> > >--
> > >Norman Levin
> > >vm/dynAmIX inc.

--
Dmitry Gordov
Instructor
IBM Learning services
Russia
 
 
 
Top

1. Apache 1.3.3/stronghold 2.4.1: non-Sec page -> Sec page problem

Hi,

I have a plain Apache 1.3.3 server running on port 80 of a machine here.
It has *Many* v-hosts tied to it, all on port 80, with their own IP
numbers.

On the same machine there is a Stronghold 2.4.1 server that runs only on
port 443.  There are no v-hosts associated with this server.

The Apache server runs on the _same_ IP address as the Stronghold server
but the port numbers are different.

I am seeing this problem with an OLD unix version of Netscape 3.0 Gold.
I have seen the behaviour with a very recent version of Internet
Exporer.

The problem:

There is a web page under a v-host on the Apache server.  It has a link
from it to a page on the Stronghold server.  I can view the document
source and the URL of the link is correct.

When I try to follow the link I get the standard "You are accessing
a secure doc" message but the "key" stays broken and there is no blue
border.  I click on "Ok" and have the page.  Looking at the "Location"
of the page it should read

        https://my.securehost.com/dir1/dir2/page.html

instead it says

        http://the.v-host.com/dir/page.html

If you type in the "https" address by hand then you go to the secure
page but not if you follow the link from the page.

The Stronghold server is getting the request because it is being logged
in that server's access_log -- however every instance is being tagged as
Status 304.

I have no redirects set up in either server config.  I do not have any
re-write rules that I am aware of.

Any ideas?

Dean...K...
--
Dean Karres               | http://www2.southwind.net/~karres

Southwind Internet Access | Programmer / Systems Administrator
Wichita, KS               | <Troll 2nd Class /w Clusters>

2. Partitioning

3. Compaq Proliant 2500: Boot problems

4. (no subject)

5. stealth video 2500 problem

6. show nr_running and nr_iowait_tasks

7. problem w/ Canopus 2500 TNT using NVidia X Server

8. SMA EtherPower 8432 BT

9. Problems with printer HP 2500 CM

10. problems with AT-2500 (8139.o)

11. Weird Network problem on Compaq 2500 /w Solaris 2.5

12. problems with diamond 2500 and 3.3x XF86

13. Diamond stealth 2500 problems


All times are UTC
Board index