Re: [Iometer-user] Linux numbers question...

[Ian <io...@ze...>]

Ah, okay the test that ended at 128 outstanding i/o's did show a differen=
ce:
Device:    rrqm/s wrqm/s   r/s   w/s  rsec/s  wsec/s    rkB/s    wkB/s
avgrq-sz avgqu-sz   await  svctm  %util
sda          0.00   2.02 102.02 29.29 2044.44 1155.56  1022.22   577.78  =
=20
24.37     1.04    7.88   7.72 101.41

average queue size was 24.  :)

I'm going to see if I can't graph the stuff I got from iostat to see if
the other numbers match up (where applicable)

> I recently ran a series of tests on a remote linux server and had some
> questions related to the tests.
>
> First off, here is the access list settings:
> File Server Access Pattern Linux,0
> 'size,% of size,% reads,% random,delay,burst,align,reply
> 4096,80,80,100,0,1,0,0
> 8192,2,80,100,0,1,0,0
> 16384,4,80,100,0,1,0,0
> 32768,4,80,100,0,1,0,0
> 65536,10,80,100,0,1,0,0
>
> Under Test Setup I created an oustanding i/o range of 1 to 16 Power 2
> (Exponential).
>
> This was on an ext3 partition. (Prepared)
>
> During the entire test, I had 'iostat -xd  1' running on the linux serv=
er
> and thoughout the test, the numbers changed little:
> (At the beginning of the test)
> Device:,rrqm/s,wrqm/s,r/s,w/s,rsec/s,wsec/s,rkB/s,wkB/s,avgrq-sz,avgqu-=
sz,await,svctm,%util
> sda,0.00,10.10,183.84,21.21,2763.64,719.19,1381.82,359.60,16.99,1.13,5.=
48,4.93,101.01
>
> (At the end of the test)
> Device:,rrqm/s,wrqm/s,r/s,w/s,rsec/s,wsec/s,rkB/s,wkB/s,avgrq-sz,avgqu-=
sz,await,svctm,%util
> sda,0.00,0.00,104.04,25.25,2125.25,444.44,1062.63,222.22,19.88,1.01,7.8=
1,7.84,101.41
>
> The number I've got a question about is the avgqu-sz (Average Queue Siz=
e)
> which always stayed at 1.  I had expected that to climb as the i/o's
> stepped  up higher, but I'm proabably not understanding what these numb=
ers
> truely represent.  It seemed that while the drive was being utilized to
> it's fullest (It always stayed at 100%), I was unable to get the system=
 to
> "hurt" so to speak by causing large queues and wait times, nor did I se=
e
> any large spikes in data per second (as far as iostat reported). I'm go=
ing
> to run the test again, this time with a range if 1 to 128 to see if tha=
t
> has any difference.
>
> (Here's the definitions for each of those numbers)
> rrqm/s: The number of read requests merged per second that were queued =
to
> the device.
>
> wrqm/s: The number of write requests merged per second that were queued=
 to
> the device.
>
> r/s: The number of read requests that were issued to the device per
> second.
>
> w/s: The number of write requests that were issued to the device per
> second.
>
> rsec/s: The number of sectors read from the device per second.
>
> wsec/s: The number of sectors written to the device per second.
>
> rkB/s: The number of kilobytes read from the device per second.
>
> wkB/s: The number of kilobytes written to the device per second.
>
> avgrq-sz: The average size (in sectors) of the requests that were issue=
d
> to the device.
>
> avgqu-sz: The average queue length of the requests that were issued to =
the
> device.
>
> await: The  average  time (in milliseconds) for I/O requests issued to =
the
> device to be served.  This includes the time spent by the requests in
> queue and the time spent servicing them.
>
> svctm: The  average  service  time  (in  milliseconds) for I/O requests
> that were issued to the device.
>
> %util:  Percentage of CPU time during which I/O requests were issued to
> the  device  (bandwidth utilization  for the device). Device saturation
> occurs when this value is close to 100%.
>
>
>
> -----------------------------------------------------------------------=
--
> Using Tomcat but need to do more? Need to support web services, securit=
y?
> Get stuff done quickly with pre-integrated technology to make your job
> easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geron=
imo
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
> _______________________________________________
> Iometer-user mailing list
> Iom...@li...
> https://lists.sourceforge.net/lists/listinfo/iometer-user
>