Re: Request: I/O request recording

From: Andrew Morton (akpm_at_osdl.org)
Date: 01/25/04

  • Next message: Andrew Morton: "Re: [PATCH][2.6-mm] Fix CONFIG_SMT oops on UP"
    Date:	Sat, 24 Jan 2004 15:35:51 -0800
    To: Felix von Leitner <felix-kernel@fefe.de>
    
    

    Felix von Leitner <felix-kernel@fefe.de> wrote:
    >
    > I would like to have a user space program that I could run while I cold
    > start KDE. The program would then record which I/O pages were read in
    > which order. The output of that program could then be used to pre-cache
    > all those pages, but in an order that reduces disk head movement.
    > Demand Loading unfortunately produces lots of random page I/O scattered
    > all over the disk.

    I wrote a similar thing in September of 2001. What you do is:

    - Reboot the system, wait until everything is steady-state (eg: X has
      started, applications are loaded).

    - Load a kernel module which dumps the current contents of the pagecache
      (filename/offset-into-file) into a file.

      (The kernel module writes to modprobe's stdout, so you just do

            modprobe fboot-dump > /tmp/fboot-dump.out

       I'm very proud of this.)

    - Post-process the resulting output into a database which is used on the
      next reboot.

    - reboot

    - This time a userspace application cuts in real early and reads the
      database and preloads all the pagecache using "optimal" I/O patterns so
      that everything which you will need in the subsequent boot is already in
      memory.

    So it's all an attempt to optimise the boot-time I/O patterns. It was
    pretty much a waste of time, gaining only 10% or so, from memory. You
    could get just as much or more speedup from simply launching all the
    initscripts in parallel, although this did tend to break stuff.

    Anyway, the code's ancient but might provide some ideas:

            http://www.zip.com.au/~akpm/linux/fboot.tar.gz

    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/


  • Next message: Andrew Morton: "Re: [PATCH][2.6-mm] Fix CONFIG_SMT oops on UP"

    Relevant Pages

    • Re: vm_dirty_ratio seems a bit large.
      ... > the machine that first reported this is running with 384GB of memory. ... All that dirty pagecache allows us to completely elide I/O when overwrites ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: Request: I/O request recording
      ... > next reboot. ... > that everything which you will need in the subsequent boot is already in ... > pretty much a waste of time, gaining only 10% or so, from memory. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: 2.6.12-rc6-mm1 & 2K lun testing
      ... I've seen this before on a system with lots of luns, lots of memory ... was triggered (58GiB of RAM in the elevator queues waiting for I/O ... SGI Australian Software Group ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Page cache write performance issue
      ... I've noticed the following performance regression from ... 1GB), writes are in 1K chunks, most of memory is free, ... cos we're initiating I/O almost from the start. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: Something corrupts raid5 disks slightly during reboot
      ... corruption after reboot and perfect work inbetween. ... Any theories on how faulty memory could actually cause something like this? ... A bad spot in memory on an area where the bios code is cached, ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)