[Disksim-users] Disksim-users Digest, Vol 91, Issue 4

[信宏陳]

hello, dear Yipkei Kwok

is that possible that your storage constantly finish your IO request (30
IOPS ?)

so the pending requests does not match the request rate


2013/7/12 <disksim-users-request at ece.cmu.edu>

> Send Disksim-users mailing list submissions to
>         disksim-users at ece.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://sos.ece.cmu.edu/mailman/listinfo/disksim-users
> or, via email, send a message with subject or body 'help' to
>         disksim-users-request at ece.cmu.edu
>
> You can reach the person managing the list at
>         disksim-users-owner at ece.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Disksim-users digest..."
>
>
> Today's Topics:
>
>    1. Problem: No. of in-flight requests bounded by 20 (Yipkei Kwok)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Wed, 10 Jul 2013 20:30:23 -0600
> From: Yipkei Kwok <ykwok2 at miners.utep.edu>
> Subject: [Disksim-users] Problem: No. of in-flight requests bounded by
>         20
> To: Disksim Users Mailing List <disksim-users at ece.cmu.edu>
> Message-ID:
>         <
> CAF75uxwp0y1Obdb_Nxpsa4ubD5J2GRZBPxkSXFZ0gYc3Kv+twQ at mail.gmail.com>
> Content-Type: text/plain; charset="ISO-8859-1"
>
> Hi,
>
> I want to use synthetic workload generators to simulate a scenario
> where 50 single-thread independent processes are concurrently
> accessing a single-disk storage system. All I/O requests are
> synchronous. I expect that the number of pending request in the entire
> I/O hierarchy (I/O driver + controller + disk) to be 50. However, I
> found that the number of pending requests drops from 50, at the
> beginning, to 20, later in the simulation.
>
> Here is the summary of the DiskSim parv file
> (http://utminers.utep.edu/ykwok2/tmp/20130710-0.parv.txt)
> * I/O driver: FCFS
> * single disk: MAXTOR_ATLAS10K5_300 (disk model provided by Anjo at
> http://ovahldy.blogspot.com/2011/03/recent-diskmodels-for-disksim.html)
> * Think time from call to request = 0.0
> * Think time from request to return = 0.0
> * number of synthetic workload generators = 50
> * System call/return with each request = 1 (i.e., each request occurs
> within the context of a system call)
> Note: I gave a try to set "System call/return with each request" to 0
> but I encountered the same problem.
> * For each generator, Probability of time-critical request = 1.0
>
> This is how I measure the number of pending requests in the I/O
> hierarchy. I keep a counter in the simulator. When a request arrives
> the I/O driver, iodriver_request() increments the counter. When a
> request is serviced, iodriver_access_complete() decrements the
> counter.
>
> The plot of the number of pending requests vs. time is here
> (http://utminers.utep.edu/ykwok2/tmp/Slide1.PNG). I repeated the
> experiment with 35, 20, and 15 generators (all results are shown on
> the same slide). The number of pending requests with 35 generators,
> too, drops to 20, while the numbers of pending requests with 20 and 15
> generators remain constant.
>
> I want the number of pending requests to be 50 when 50 generators are
> used. Why is the number of pending requests dropping?
>
> Please kindly help. Thank you very much.
>
> Regards,
> Yipkei
>
> --
> ********************************************
> Yipkei Kwok
> Ph.D. Candidate
> HiPerSys Lab
> Department of Computer Science
> The University of Texas at El Paso
> Phone: 915 747 6433 (O)
> E-mail: ykwok2 at miners dot utep dot edu
> ********************************************
>
>
>
> ------------------------------
>
> _______________________________________________
> Disksim-users mailing list
> Disksim-users at ece.cmu.edu
> https://sos.ece.cmu.edu/mailman/listinfo/disksim-users
>
>
> End of Disksim-users Digest, Vol 91, Issue 4
> ********************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.andrew.cmu.edu/pipermail/disksim-users/attachments/20130712/f0943233/attachment.html>